Indonesian Journal of Electrical Engineering and Computer Science

Vol. 32, No. 3, December 2023, pp. 1712~1723

ISSN: 2502-4752, DOI: 10.11591/ijeecs.v32.i3.pp1712-1723  1712

Poultry disease early detection methods using deep learning

technology

Liu Yajie1,2, Md Gapar Md Johar3, Asif Iqbal Hajamydeen4

1
School of Graduate Studies, Management and Science University, Shah Alam, Malaysia
2
School of Information Engineering, Henan Vocational University of Science and Technology, Zhoukou, China
3
Software Engineering and Digital Innovation Centre, Management and Science University, Shah Alam, Malaysia
4
Faculty of Information Sciences and Engineering, Management and Science University, Shah Alam, Malaysia

Article Info ABSTRACT

Article history: Poultry production is a pivotal contributor to global economic growth, playing
a central role in promoting human ecosystem sustainability. It offers
Received Sep 7, 2023 affordable and readily accessible protein sources, encompassing meat, eggs,
Revised Sep 27, 2023 and other by-products. Beyond its direct nutritional benefits, poultry
Accepted Oct 3, 2023 production enhances household income, bolsters food security, and aids in
poverty reduction, making it integral to worldwide economic advancement.
However, as the global population surges, so does the demand for poultry
Keywords: meat and eggs. Concurrently, poultry disease management emerges as a
paramount challenge, leading to significant threats to food security and
Classification economic stability. Leveraging cutting-edge technology offers promising
Clinical symptoms avenues to devise strategies that not only bolster farm profitability but also
Deep learning mitigate environmental impacts and foster the well-being of both animals and
Early detection humans. This study systematically reviews the latest literature concerning
Early warning poultry disease diagnosis based on deep learning techniques, elucidating the
Intelligent management clinical manifestations associated with various ailments. The analysis
Poultry disease indicates that emerging technological solutions, especially image processing
and deep learning (DL), substantially outperform conventional manual
inspection methods in early disease detection and warning in the poultry
sector. Such innovations underscore their potential for revolutionizing poultry
health management and disease mitigation.
This is an open access article under the CC BY-SA license.

Corresponding Author:
Liu Yajie
School of Graduate Studies, Management and Science University
Shah Alam, Selangor, Malaysia
Email: [email protected]

1. INTRODUCTION
The poultry industry plays a critical role in global economic development. It contributes to
socioeconomic growth by providing people with affluent protein and nutrition [1]. The development of
agriculture is the foundation of a country's economic progress. Humans consume a substantial amount of meat
and eggs produced by the poultry industry [2]. Worldwide affluence and population growth fuel food demand,
resulting in increased poultry production in numerous nations. By 2030, it is estimated that the global
population will reach 8.6 billion, posing a significant challenge to adequate meat production and provision [3].
It is estimated that 40% of poultry meat and egg demand will increase by 2050 [4]. Poultry meat and eggs are
essential to human life because they are a cost-effective and valuable protein source in the daily diet. Poultry
farming is also a direct source of family income for many local farmers. Due to its small size and low
investment, the small-scale poultry feeding industry is a reliable source of employment and family income for

Journal homepage: http://ijeecs.iaescore.com

Indonesian J Elec Eng & Comp Sci ISSN: 2502-4752  1713

a variety of households [5]. It is imperative to maintain sustainable food production and livelihoods at multiple
scales to achieve sustainable development goals. While a substantial industry opportunity arises from
increasing poultry demand, concurrent challenges persist. As climate change, land erosion, biodiversity loss,
and biosecurity threats increase, concerns about human health and animal protein have intensified and are being
focused on. As long-term environmental changes occur, new demands, challenges, and pressures are arising as
well [6]. The poultry farming industry is susceptible to a number of diseases, including avian influenza,
Salmonellosis, fowl cholera, Newcastle, and Coccidiosis [7]–[13]. Figure 1 describes poultry diseases'
transmission routes and clinical manifestations.

Figure 1. The routes of poultry diseases transmission and clinical manifestations [2]

Influenza affects not only the quantity of poultry meat and eggs produced but also their quality, which
are typically significant effects of avian influenza. Additionally, birds have been observed to suffer from a high
rate of morbidity and mortality. It is common for avian influenza to be transmitted through feces, the mouth,
or aerosol [8]. Several clinical signs may be observed, such as dehydration, loss of appetite, fever, depression,
decreased intake of food and water, a decrease in response to stimuli, wing paralysis, torticollis, and
tremors [9]. So, it is paramount to adopt a mechanism for early avian influenza infection identification before
outbreaks. Salmonellosis is another one of the most widespread bacterial zoonotic diseases, with around
155,000 deaths observed around the world. In the poultry industry, it is one of the most prevalent diseases,
which is caused by varieties of Salmonella. Salmonellosis could be transmitted not only through the horizontal
fecal-oral pathway but via vertically embrocated eggs as well. It is common for mature poultry to exhibit
decreased egg production, fertility, and hatchability, as well as anorexia, diarrhea with white or yellow mucus,
and watery mucosa [10]. Salmonellosis could be spread in poultry farms through infected sick chickens,
contaminated equipment, water and feed, domestic animals, and the surrounding environment as well. Fowl
cholera is a kind of bacterial-infected disease in poultry farming. Fowl cholera is a septicemic disease affecting
wild and domestic birds. Mature poultry are more susceptible to infection than young ones.
The transmission route for poultry cholera can be oral, nasal discharge, or feces. Some typical clinical
symptoms include nasal discharge, respiratory rales, and coughing [1]. At the same time, as the joints continue
to swell and deform, some inflammatory substances leak out, ultimately leading to necrosis [11]. Apart from
that, clinical symptoms like ruffled feathers, mucous discharge from the mouth and nose, diarrhea, and general
depression are addressed as well [12]. Newcastle disease is a viral poultry disease. A group of viruses closely
related to the avian paramyxovirus type 1 serotype causes Newcastle disease. In poultry farming, it is one of
the most feared diseases and one of the leading causes of mortality. Some clinical signs, such as respiratory
and nervous symptoms. Oral is the most common transmission route, but respiratory and conjunctival modes
may also be included [13]. Newcastle disease is characterized by a variety of major symptoms including
coughing, gasping, greenish diarrhea, sneezing, cyanosis of the comb and wattle, corticoids, paralysis (wings
and legs), tremors, and a twisted neck. Ducks and geese have stronger immune systems than chickens and
turkeys against this disease. As with pathogenic avian influenza, Newcastle disease presents very similar
clinical symptoms. Early isolation of virus-infected poultry is the most effective countermeasure [14].
Coccidiosis is a parasitic disease in poultry farming. A protozoal parasite causes this disease, making it one of
the most prevalent diseases in poultry. Some clinical symptoms include poor growth, feed conversion, and
even death under severe conditions. Furthermore, parasites could decrease herd immunity to other diseases [1].
Most poultry diseases are characterized by ruffled feathers, depression, panting, diarrhea (watery or
bloody), drooling saliva, coughing, head and eye swelling, curling of the head and neck, and decreased egg
production. Based on the above clinical symptoms, sick poultry could be diagnosed by vocalization, body
temperature, feces, and daily behaviors, allowing for health evaluation [2]. Handling sick poultry cautiously
Poultry disease early detection methods using deep learning technology (Liu Yajie)
1714  ISSN: 2502-4752

and keeping normal poultry away from potentially ill ones in time is critical in poultry farming, especially in
large-scale farming. So, timely isolating the symptoms-appearing chickens could prevent healthy chickens
from getting infected [15].
To combat disease in poultry farms, prevention, vaccines, and medications are the commonly used
mechanisms to enhance health conditions and overall production [16]. Poultry industry production could be
enhanced with the appropriate strategies for husbandry management and clean feeding environments.
Traditional methods involve veterinarians conducting manual observation or biochemical testing, which are
time-consuming and labor-intensive. Biochemical tests are sensitive but expensive, and manual inspection
could result in incorrect results. Moreover, traditional inspection methods may miss the appropriate time for
disease treatment, especially in cases of severe infectious diseases. With the modern technology development,
internet of things (IoT) applications, video and image processing, classification capabilities, and smart poultry
management have been emphasized in recent years [17]–[20]. The low cost of computing resources and
common algorithms make contemporary technology an indispensable tool for monitoring vast farms and
increasing production [21]. Advanced technological solutions are crucial for poultry farm health management,
as poultry meat and eggs are the largest protein sources [22]. Disease monitoring and early detection are
essential for decreasing poultry morbidity and mortality and boosting production yields. By deploying modern
technologies, sick poultry early identification and warning could be automatically and consistently controlled.
It aims to reduce infection rates by isolating sick poultry in a timely manner.
Nowadays, deep learning technology has the capability of self-learning to analyze images
automatically, which enables the constructed deep learning model to facilitate the analysis and management of
the poultry farming industry, especially in sick poultry early detection. Powered artificial intelligence (AI) and
deep learning technologies could be deployed into the data analysis process to analyze, predict, and inform
end-users of abnormal conditions to reduce the spread of poultry diseases and ensure biosecurity. The
application of big data presents an unprecedented opportunity in the development of tools that will optimize
farm profitability, reduce environmental impact, and increase the health and welfare of animals and
humans [23]. To improve animal health and reduce losses, it is imperative to detect abnormalities in poultry
and issue early warnings of infectious diseases. Nevertheless, inadequate methods could result in decreased
productivity and extensive mortality. Symptom detection technologies could continuously, noninvasively, and
automatically monitor the health conditions of laying hens and broilers, which could aid in making early disease
warning decisions. However, clinical symptom-based monitoring systems for on-farm disease detection have
not been fully implemented [2].
All the references selected for this project are collected from the Google Scholar and Web of Science
databases. A collection of searching keywords is used to find suitable references: "Poultry/chicken
disease/behavior/posture detection/monitoring/warning", "Deep/machine learning", "Smart/intelligent poultry
farm" and similar words. There are several criteria set to include the specific reference: i) the study mentions
at least one method to diagnose the disease based on poultry disease clinics in detail; ii) the proposed
methodology is related to machine learning or deep learning algorithms; and iii) the proposed system is
constructed and experimented with. Based on the above conditions, the related references are selected. A
review and summary of these references are used to provide an in-depth understanding of the application based
on deep learning techniques in the early diagnosis of poultry diseases, making a significant contribution to this
research.

2. LITERATURE REVIEW
Poultry's physiological traits include rich information about their environmental condition, emotional
state, and health status. This information can be used to monitor the welfare of poultry and inform decisions
about husbandry practices. Studies have shown that physiological traits could be used to detect changes in the
environment and animal health. In poultry, some common physiological traits like body temperature,
vocalization, and feces are associated with different diseases [24]. So, these factors are utilized and measured
to detect and diagnose diseases in their early stages to avoid potential risks.

2.1. Abnormal body temperature

Poultry is a homeothermic animal that creates and distributes heat to keep core and skin surface
temperatures constant. This is called homeothermy, and it assists poultry in regulating its metabolism and
improving its physiological functions. Homeothermy improves the bird's capacity to obtain food and thrive in
a variety of situations [25]. The poultry industry faces significant economic losses due to heat stress (HS),
which affects production performance, body temperature, and immune responses. One of the most important
ways to identify HS possessions in poultry is to monitor their body temperature during rearing and take action
in time [26]. The temperature would change according to different pathological or stress conditions, such as

Indonesian J Elec Eng & Comp Sci, Vol. 32, No. 3, December 2023: 1712-1723
Indonesian J Elec Eng & Comp Sci ISSN: 2502-4752  1715

disease infection or an instinctive response. Therefore, temperature change could be one significant factor that
indicates early warning for sick poultry. High temperatures decrease feed efficiency, body weight, feed
consumption, and egg production, increasing mortality and pathological damage [27]. The temperature of the
animal body is intimately associated with metabolism and living activities, and it reflects their physiological
and health status. Infrared thermography (IRT) technology is a type of non-invasive monitoring technique used
to assess animal health and physiological changes. IRT has been used in animal temperature detection for
disease detection, extreme thermoregulation, and estrus detection [28]. It is possible to determine poultry
surface temperatures using infrared cameras. This is done by creating images displayed in different colors
corresponding to different temperatures using imaging techniques. Chicken body temperature can be measured
by applying this technology after changes in diet, stress levels, and environmental conditions [23].
Noh et al. [29] developed an innovative system for real-time surveillance of infected poultry before disease
manifestation. It showed if it would be possible to employ thermal imaging to find changes in the surface
temperature of ducks and chickens as an early sign of highly pathogenic avian influenza (HPAI) infection.
Thermal camera footage was used to detect H5N1 infections in chickens and ducks. The paper suggested using
thermal imaging cameras in livestock to detect early signs of the avian influenza virus. It also suggested a
maximum change in surface temperature that should be considered when deciding to kill infected birds.
Chuang et al. [30] proposed a goose surface temperature monitoring system for commercial poultry houses.
It developed a deep learning model to automatically identify geese in visible images and automatically
determine individual goose surface temperatures based on infrared (IR) thermal images. This model used
convolutional neural networks (CNNs) to find geese in visible images and a second CNN to figure out the
surface temperature by looking at the IR thermal images of the geese. By using both visible images and infrared
data, the model can accurately classify the geese and figure out their surface temperature.

2.2. Abnormal vocalization

Recognizing and diagnosing illnesses in chickens is critical for the health and well-being of poultry
flocks. Visual observation and physical examination have traditionally been the primary methods used to
identify sick chickens. However, new research suggests that abnormal vocalizations emitted by sick chickens
could serve as a potential indicator of their health status. Figure 2 describes the different time domain
characteristics of chickens: healthy, infected Newcastle, bronchitis virus, and avian influenza. Another way to
detect illnesses in chickens is by analyzing their vocalizations for abnormal patterns and characteristics using
voice recognition technology [31].

Figure 2. Four time domain signal characteristics [32]

Abnormal voice detection is another way to identify sick poultry in the early stages. This novel
approach has the potential to supplement existing diagnostic methods by providing a non-invasive and efficient
method of early illness detection on poultry farms. Quintana et al. [33] have developed a hybrid solar-powered
chicken disease monitoring system using decision tree models for disease identification and verification. The
system used visual and acoustic inputs from 15 chicken samples, identifying six symptoms. After 72 hours of
continuous monitoring, the learning model achieved 84.6% accuracy in classifying diseased chickens when
visual imagery was considered, and 86.1% accuracy when audio inputs were provided. Li et al. [34] presented
a sex detection system based on chick calls for poultry breeding. This system aimed to achieve chick-call
classification and sex detection using the proposed deep learning methods. The experiment studied three
different chick breeds and used a short-time and zero-crossing rate to identify the chick call endpoints in audio.

Poultry disease early detection methods using deep learning technology (Liu Yajie)
1716  ISSN: 2502-4752

The results showed that the ResNet-50 deep neural network (DNN) had 83% highest test accuracy for
three-yellow chicks, 76.8% for native chicks, and 66.56% for flaxen-yellow chick calls. The gated recurrent
unit (GRU) and convolutional recurrent neural network (CRNN) networks achieved the highest sex
identification accuracy of 90% and 80%, respectively. Newcastle disease is a prevalent poultry disease
affecting health and production. A novel ResNet50-based system, the deep poultry vocalization network
(DPVN), was proposed for early identification using poultry vocalization. To reduce the influence of noise on
the signal, the method combined multi-window spectral subtraction and high-pass filtering. The proposed
identification system attained an average accuracy of 91.06% for infected chicks within the first, second, third,
and fourth days. A valuable benefit of this method was the improvement of animal welfare and poultry
production through automatic monitoring [35]. Jakovljević et al. [36] introduced an audio-based system to
monitor broiler chicken stress. It was based on audio signals from the first few weeks in a chicken's life to
detect stress. The system was developed to monitor the sound of birds' vocalizations during their early life
stages. By monitoring these sounds, the system could accurately detect stress and could be used to improve the
conditions in which chickens are raised. It showed that pre-recorded chicken sounds could be used with
different classifiers to figure out if a chicken was stressed or not. Each classifier in the system classes adult
chickens into a different category depending on the audio data. Four classifiers were tested at the 1,000 ms
frame level, and the accuracy of these classifiers varies from 63% to 83%, depending on the age group. Wang
et al. [37] proposed an effective audio-based system for automatically identifying and recognizing the different
types of chick vocalizations, including begging, peeping, and chirping.
The system used a deep learning algorithm to analyze audio recordings of chick vocalizations and
identify specific features that are associated with each type of vocalization. This allows the system to accurately
identify and recognize the different types of vocalizations. A new feature extraction method was used based
on joint time-frequency scattering (JTFS) transformations; varying chick calls can be distinguished 10% more
accurately. Carpentier et al. [38] developed an algorithm to detect chicken sneezing sounds in noisy
environments. The algorithm was developed using a dataset of 763 sneezes from 51 chickens. The algorithm
categorized the sounds as sneeze or non-sneeze, with 66.7% sensitivity and 88.4% precision. This work marked
the first step towards an automated, sound-based detection system for poultry health. Huang et al. [39]
presented a new audio analysis-based detection method for the early detection of avian influenza using chicken
sound and ambient noise. The extracted sound is processed using Mel-frequency cepstral coefficients (MFCC)
to differentiate between healthy and infected chickens. The method's accuracy ranges from 84% to 90%,
demonstrating the potential for efficient detection in large-scale poultry farming.

2.3. Abnormal dropping

When poultry are infected with pathogens, their droppings may appear differently. The characteristics
of chicken excrement are different between healthy and sick chickens. Different from typical healthy feces.
Diarrhea that is liquid or watery feces that are orange or red (bloody), black, dark green, yellow, watery white,
foamy, oily, or contain worms can be an indication of diseases or parasites [40]. Figure 3 describes the feces
traits related to corresponding chicken diseases. So, chicken droppings are one of the most significant signals
to indicate their health conditions.

Figure 3. Feces characteristics associated with different chicken diseases [41]

Indonesian J Elec Eng & Comp Sci, Vol. 32, No. 3, December 2023: 1712-1723
Indonesian J Elec Eng & Comp Sci ISSN: 2502-4752  1717

The excretion status of livestock and poultry is crucial for monitoring their health. Inspectors currently
make preliminary judgments by observing the color and traits of feces on the manure belt. The identification
of feces from poultry is critical for food safety and disease prevention. Fluorescence imaging could detect
excrement, but it requires skill. Using fluorescence imaging and deep learning approaches, the study recognized
disease-associated feces types from feces photos. EfficientNet-B0 had an accuracy of 97.32% in segmenting
feces, while U-Net had an accuracy of 89.34% [42]. Zhu and Zhou [43] proposed a machine vision-based
chicken manure image recognition method for online monitoring. The method preprocessed the collected
images, made preliminary judgments about abnormal manure, and analyzed grayscale characteristics to
determine normality. The method was effective for monitoring images of abnormal chicken droppings and can
initially determine the health condition of chickens. Fecal imaging is essential for determining the health of
poultry, but producers frequently struggle with disease diagnosis. Aworinde et al. [44] developed a dataset of
images of healthy and ill feces from poultry farms in Nigeria. The dataset comprises 14,618 labeled images
that can be utilized in machine learning models and computer vision applications. This dataset was designed
to assist farmers and agricultural extension agents in managing poultry farms, minimizing losses, maximizing
profit, and maximizing protein sources.
Mbelwa et al. [45] proposed a chicken feces classification system using a CNN deep learning solution.
The XceptionNet model outperforms other models in all metrics, with a 94% validation accuracy using
pretraining. The fully trained CNN comes in second, while the pre-trained XceptionNet method has the highest
prediction accuracy, making it suitable for chicken disease detection applications. Due to late diagnoses and a
dearth of credible specialists, diagnostic methods for chickens, such as oocyte count, virus detection, and
polymerase chain reaction (PCR), are frequently inadequate. Suthagar et al. [46] proposed a model for the early
detection and classification of poultry diseases using a database based on feces. The dataset consists of
6,812 images divided into four categories: healthy chicken, Coccidiosis, Salmonella, and Newcastle. Deep
learning techniques, such as pre-trained DenseNet, Inception, and MobileNet, accurately predict chicken feces
with 97% accuracy, making them appropriate for use in poultry diagnostic applications. Widyawati and
Gunawan [47] presented a study conducted to detect early-sick chickens on a real poultry farm in Indonesia
using the YOLOv5 algorithm. This study was carried out through the analysis of chicken feces' image features.
The results of this research achieved 89.2% accuracy.

2.4. Early disease detection through behavioral characteristics

Early disease detection is a critical aspect of effective poultry health management, as it enables prompt
intervention and inhibits the spread of illness within flocks. While traditional diagnostic methods rely on
physiological and laboratory-based indicators, emerging research suggests that behavioral characteristics could
serve as an additional factor for identifying sick poultry. By closely monitoring their behavior and recognizing
abnormal patterns, such as changes in feeding habits, reduced activity levels, or abnormal body pose
characteristics, it may be possible to detect early indicators of illness in poultry [48]. Figure 4 shows the
behaviors of normal and sick chickens [49].
This innovative approach of using behavior identification as a tool for early disease detection holds
significant potential for enhancing poultry health surveillance and allowing for prompt intervention, ultimately
safeguarding the welfare and productivity of chicken populations. Animal behavioral analysis is becoming
increasingly important for farm animal welfare, health, efficiency, and sustainable environments. Behavioral
analysis can help identify poultry diseases through accurate pose estimation, which can help farmers diagnose
or isolate sick poultry. Hilmi et al. [50] studied the use of deep convolutional networks (DeepCNN) to
accurately detect poultry's body key points in videos, allowing for the development of an automated poultry
health classifier. The paper explained the data gathering and tuning and compared the best available DeepCNN
using DeepLabCut, a pose estimation toolbox. Poultry pose estimation is critical for assessing abnormal
behavior and predicting sickness in poultry.
Fang et al. [51] developed a DNN approach and compared it with other techniques for estimating the
posture of individual chickens. The approach has a high precision of 95%, facilitating abnormal behavior
detection in poultry. Fang et al. [52] used DNN pose estimation and a Naive Bayesian model (NBM) to analyze
broiler chicken behavior. The method identified chickens in walking, standing, eating, running, preening, and
resting states, as shown in Figure 5, with a precision of 0.7511 for standing, 0.5135 for walking, 0.6270 for
running, 0.9361 for eating, 0.9623 for resting, and 0.9258 for preening.
This non-invasive method offered valuable insights for future behavior analysis to identify sick
chickens in broiler chicken farming. Poultry behavior is critical for health and well-being, and grill producers
must detect lameness early. Using video data, Nasiri et al. [53] built a position estimation-based model to detect
lameness in broilers. DeepCNN were used to recognize seven important spots on walking broilers, which were
then fed into a long short-term memory (LSTM) model. The model has 95% overall classification accuracy
and 97.5% average per-class classification accuracy, making it an effective and non-invasive tool for chicken
farms. Xie and Chang [54] presented a method for classifying broiler behaviors by using object detection and
Poultry disease early detection methods using deep learning technology (Liu Yajie)
1718  ISSN: 2502-4752

recurrent-based DNNs. The YoLov4 object detection model was trained and used to detect five pre-defined parts
of a broiler chicken, and then the corresponding chicken skeleton was constructed from these identified parts.

Figure 4. Samples of sick and normal chickens. From left to right, the labels are normal chickens, chickens
with chicken head disease, and chickens with ILT, Newscalte, and Marek diseases [49]

Figure 5. Typical chicken behaviors [52]

The angle between backbone fulcrum vectors was extracted. Six broiler behaviors were detected using
a time-series-based LSTM network. The scheme was validated in an outdoor environment, with average
precision, recall, and F1-score of 82%, 81%, and 81%, respectively. Using broilers as an example, Zhang and
Chen [55] proposed an autonomous detection method for unwell chickens. To improve the network structure
and adapt to varied recognition situations, the system employed the ResNet residual network. The model's
identification rate on the test set improved by 2.1% after tens of thousands of recurrent training sessions,
resulting in a 93.7% recognition rate. This intelligent management platform focused on remote monitoring of
cattle and poultry breeding conditions, dealing with behavioral physiology, and tracking production
performance. A camera-based system was developed to monitor flocks and detect injuries using neural
networks.

Indonesian J Elec Eng & Comp Sci, Vol. 32, No. 3, December 2023: 1712-1723
Indonesian J Elec Eng & Comp Sci ISSN: 2502-4752  1719

A preliminary study used manually annotated 244 images of turkeys to train a neural network called
HRNet-W48 for injury detection. Seven turkey key points were defined. The proposed model achieved an
average accuracy of 73.5%, and showed positive influences on turkey management, allowing for clear
differentiation between individual animals even in crowded situations [56]. Bakar et al. [57] developed a
supervised machine learning algorithm for early detection of bacteria- or virus-infected chickens using the
International Commission on Illumination (CIE) XYZ color space. The algorithm uses a logistic regression
model to classify chickens, revealing 100% sensitivity and 83% specificity. The comb chromaticity of infected
chickens’ changes from red to green, contributing to the development of modern technology in agriculture
applications. The detection method feature serves as an indicator for detecting bacteria or virus-infected
chickens, contributing to the advancement of modern technology in agriculture applications shown in Table 1,
it is summarizes various early detection methods associated with poultry diseases.

Table 1. Summarization of poultry abnormality detection methods

Author Year Reference Discussion Results
Noh et al. 2021 [29] Developed a real-time surveillance Suggested using thermal
system for poultry surface temperature imaging cameras in livestock to
monitoring based on a thermal camera. detect early signs of the avian
influenza virus.
Chuang et al. 2021 [30] Proposed a goose surface temperature An infrared thermal camera
monitoring system based on a CNN combined with a CNN model
algorithm. could detect individual surface
temperature changes accurately.
Quintana et al. 2022 [33] Developed a hybrid solar-powered The proposed model achieved
chicken disease monitoring system up to 86.1% accuracy in
using decision tree models for disease classifying diseases.
identification and verification based on
various audios.
Jakovljević et al. 2019 [36] Introduced an audio-based system to The system achieved a range of
monitor broiler chicken stress using 63%–83% accuracy.
support vector machines (SVM)
algorithm.
Carpentier et al. 2019 [38] Developed an algorithm to monitor Achieved a precision of 88.4%.
chicken sneezing sounds in noisy
environments.
Huang et al. 2019 [39] Presented an audio-based detection Achieved up to 90% accuracy.
method for the early detection of avian
influenza based on MFCC method.
Gorji et al. 2022 [42] Demostration of using fluorescence Achieved up to 97% accuracy.
imaging and deep learning approaches
to recognize poultry feces.
Mbelwa et al. 2021 [45] Proposed a CNN-based solution to Achieved a 94% validation
predict chicken feces' classification. accuracy.
Suthagar et al. 2023 [46] Proposed a feces-based deep learning Achieved up to 97% accuracy.
model for the early detection and
classification of poultry diseases.
Widyawati and 2022 [47] Presented a feces-based model to Achieved 89.2% accuracy.
Gunawan detect early sick chickens using YOLO
V5 algorithm.
hilmi et al. 2022 [50] Developed ResNet-based pose Poultry pose estimation is
estimation toolbox to detect poultry's critical for assessing abnormal
body keypoints. behavior and predicting
sickness in poultry.
Fang et al. 2022 [51] Developed a DNN approach to Achieved 95% accuracy.
estimate the posture of individual
chickens.
Fang et al. 2021 [52] Developed a DNN pose estimation Achieved 78.64% average
model to analyze chicken behavior. precision.
Xie and Chang 2022 [54] Presented a method for identifying Achieved 82% accuracy.
broiler behavior based on YOLO v4
algorithm.
Gourisaria et al. 2023 [58] Proposed a CNN-based model called Achieved 94.49% accuracy.
ChicNetV6 to classify different
diseases.
Yang et al. 2023 [59] Developed a CNN-based model to Achieved 95.3% average
classify six chicken behaviors. accuracy.

Gourisaria et al. [58] utilized CNN models to classify diseases like Salmonella, Coccidiosis, Healthy,
and NewCastle disease. A six-behavioral classifier was developed to monitor natural behaviors in cage-free
Poultry disease early detection methods using deep learning technology (Liu Yajie)
1720  ISSN: 2502-4752

birds, including feeding, drinking, walking, perching, dust bathing, and nesting. The classifier achieved an
average accuracy of 95.3%, with the highest accuracy for drinking behavior in chicks (97.8%) and 92.5% for
nesting behavior. The classifier is useful for separating cage-free bird behaviors across various life periods and
environments [59]. Based on the identified behaviors, the activities of monitored poultry could be further
analyzed. For example, if walking and drinking activities were reduced, individuals would be most susceptible
to pathogens and might already be infected [60].

3. DISCUSSION
In the domain of poultry farming, effective disease control remains a paramount concern.
Conventional management methodologies, though human-resource intensive, often fail to meet the escalating
demands of contemporary poultry operations, particularly in large-scale establishments where manual
inspection approaches are subjective and labor-intensive. The integration of modern technologies, such as AI
and deep learning, CNN, in particular, have achieved outstanding results in image classification tasks. This
provides possibilities for the health management of poultry and the automatic analysis of poultry diseases. It
promises to augment the efficiency of poultry farming by facilitating automated health management and real-
time early disease detection based on clinical indicators. While the convergence of human expertise and AI in
the livestock sector does present nuanced challenges, the accrued benefits appear to significantly overshadow
these concerns. Current deep learning techniques, characterized by their inherent self-learning capabilities,
have demonstrated proficiency in diverse poultry management scenarios, notably in early disease detection.
Utilizing computer vision technology in livestock welfare and health management research streamlines
continuous poultry observation, making it more efficient. The synergy between computer vision and deep
learning is poised to transform conventional poultry farming practices, making them more modern and high-
production.
Currently, a myriad of non-invasive early poultry disease detection techniques, such as abnormal
vocal pattern analysis, dermal temperature assessment, fecal analysis, postural estimation, and behavior
monitoring, are the subjects of extensive global research. These non-invasive diagnostic methods do not cause
stressful effects on the poultry and ensure the normal habits of the poultry. Beyond mitigating large-scale
disease outbreaks, these innovations have the potential for broader applications in extensive animal-rearing
contexts, including traditional poultry farming. In this way, farmers can take precautions in advance to
minimize the impact of diseases. Despite the great potential of deep learning technology, there are still
challenges in applying it in the field of poultry farming. First, a large amount of labeled data is needed for
model training. In addition, the actual farming environment may affect the accuracy of the sensor data.
Economics also needs to be considered, i.e., whether the cost of collecting and processing the data can be offset
by the profitability of the farming industry.

4. CONCLUSION
Poultry diseases pose formidable threats to both small-scale and industrial-scale farming practices.
Conventional manual inspection methods, while foundational, have become labor-intensive and often fall short
in addressing the demands precipitated by the escalation of poultry populations and the expansion of farming
scales. Leveraging contemporary technology, it's feasible to develop a method that provides real-time, non-
invasive, and efficient health monitoring for poultry, facilitating automated disease detection and immediate
intervention alerts. Current advancements in intelligent poultry health monitoring predominantly employ
metrics such as dermal temperature variations, vocalization traits, fecal characteristics, posture, and specific
behavioral characteristics. The integration of AI and advanced sensing technologies promises to elevate
operational efficiencies and commercial prospects for farmers, while also advancing the welfare of both
animals and the humans involved. The future development trend of the poultry sector seems to be steering
toward more intelligent management and automatic health monitoring systems. A sophisticated remote
monitoring mechanism promises continuous and automated oversight, with a focus not just on the
environmental conditions but also on the holistic health of the poultry. By deploying an intelligent early
warning system for poultry diseases, there is potential to drastically diminish the reliance on manual
monitoring, enhance the overall health of poultry, streamline production processes, and boost profitability. By
capitalizing on the robust computational capabilities of deep learning, it becomes possible to undertake real-
time data analysis. Should any discrepancies or unusual patterns in health data arise, farm managers are
promptly alerted, ensuring proactive measures are instituted to thwart potential disease outbreaks. For a more
precise analysis, it is imperative to encompass a comprehensive understanding of the multifaceted attributes
associated with the disease in question.

Indonesian J Elec Eng & Comp Sci, Vol. 32, No. 3, December 2023: 1712-1723
Indonesian J Elec Eng & Comp Sci ISSN: 2502-4752  1721

REFERENCES
[1] T. A Serbessa, Y. G Geleta, and I. O Terfa, “Review on diseases and health management of poultry and swine,” International
Journal of Avian & Wildlife Biology, vol. 7, no. 1, pp. 27–38, Apr. 2023, doi: 10.15406/ijawb.2023.07.00187.
[2] P. He et al., “Research progress in the early warning of chicken diseases by monitoring clinical symptoms,” Applied Sciences
(Switzerland), vol. 12, no. 11, p. 5601, May 2022, doi: 10.3390/app12115601.
[3] OECD/FAO, OECD-FAO Agricultural Outlook 2021-2030. OECD, 2021.
[4] J. Astill, R. A. Dara, E. D. G. Fraser, B. Roberts, and S. Sharif, “Smart poultry management: smart sensors, big data, and the internet
of things,” Computers and Electronics in Agriculture, vol. 170, p. 105291, Mar. 2020, doi: 10.1016/j.compag.2020.105291.
[5] E. Abayneh, H. Goba, and M. Shurbe, “Salmonellosis prevalence and risk factors in chicken breeding farms in and around Arba
Minch town, Gamo Zone, Ethiopia,” Journal of Infection in Developing Countries, vol. 17, no. 2, pp. 226–235, Feb. 2023,
doi: 10.3855/JIDC.17553.
[6] D. Pearson, J. Gorman, and R. Aspinall, “Multiple roles for landscape ecology in future farming systems: an editorial overview,”
Land, vol. 11, no. 2, p. 288, Feb. 2022, doi: 10.3390/land11020288.
[7] F. Krammer and S. Schultz-Cherry, “We need to keep an eye on avian influenza,” Nature Reviews Immunology, vol. 23, no. 5,
pp. 267–268, May 2023, doi: 10.1038/s41577-023-00868-8.
[8] T. Adere and M. Mukaria, “The epidemiology of avian influenza and its significance for public health,” Journal Healthcare
Treatment Development, no. 31, pp. 48–58, Jan. 2023, doi: 10.55529/jhtd.31.48.58.
[9] A. N. Gkolia, “Avian bird influenza type a viruses: current situation in Europe, epidemiological considerations, prevention,
treatment and biosafety measures,” SHSSE, Int. Hellenic Univ., 2023.
[10] J. K.N. Kuria, “Salmonellosis in food and companion animals and its public health importance,” in Salmonella-Past, Present and
Future [Working Title], IntechOpen, 2023.
[11] H. Y. Qin, J. H. Xiao, J. X. Li, X. Gao, and H. B.Wang, “Climate variability and avian cholera transmission in Guangxi, China,”
Revista Brasileira de Ciencia Avicola / Brazilian Journal of Poultry Science, vol. 19, no. 2, pp. 211–220, Jun. 2017,
doi: 10.1590/1806-9061-2016-0411.
[12] K. K. Cooper, J. G. Songer, and F. A. Uzal, “Diagnosing clostridial enteric disease in poultry,” Journal of Veterinary Diagnostic
Investigation, vol. 25, no. 3, pp. 314–327, May 2013, doi: 10.1177/1040638713483468.
[13] A. V. Sharopatova, N. I. Pyzhikova, and J. A. Olentsova, “The current situation of the poultry industry and the formation of a
strategy for its sustainable development in the region,” IOP Conference Series: Earth and Environmental Science, vol. 421, no. 2,
p. 022061, Jan. 2020, doi: 10.1088/1755-1315/421/2/022061.
[14] T. Mossie and D. A. Wondie, “A compressive review on Newcastle disease virus in Ethiopia,” Journal of Veterinary Science and
Technology, vol. 13, no. 6, p. 131, 2023, doi: 10.37421/2157-7579.2022.13.131.
[15] Y. Li et al., “African swine fever in a commercial pig farm: outbreak investigation and an approach for identifying the source of
infection,” Transboundary and Emerging Diseases, vol. 67, no. 6, pp. 2564–2578, Nov. 2020, doi: 10.1111/tbed.13603.
[16] G. Bucini et al., “Connecting livestock disease dynamics to human learning and biosecurity decisions,” Frontiers in Veterinary
Science, vol. 9, no. April, Jan. 2023, doi: 10.3389/fvets.2022.1067364.
[17] F. J. Adha, M. G. Md Johar, M. H. Alkawaz, A. I. Hajamydeen, and L. Raya, “IoT based conceptual framework for monitoring
poultry farms,” in 2022 12th IEEE Symposium on Computer Applications and Industrial Electronics, ISCAIE 2022, May 2022, pp.
277–282, doi: 10.1109/ISCAIE54458.2022.9794471.
[18] N. Z. Malika, M. G. Md Johar, M. H. Alkawaz, A. I. Hajamydeen, and L. Raya, “Temperature & humidity monitoring for poultry
farms using IoT,” in 2022 12th IEEE Symposium on Computer Applications and Industrial Electronics, ISCAIE 2022, May 2022,
pp. 76–81, doi: 10.1109/ISCAIE54458.2022.9794520.
[19] N. Z. Malika, R. Ramli, M. H. Alkawaz, M. G. Md Johar, and A. I. Hajamydeen, “IoT based poultry farm temperature and humidity
monitoring systems: a case study,” in Proceeding-2021 IEEE 9th Conference on System, Process and Control, ICSPC 2021, Dec.
2021, pp. 64–69, doi: 10.1109/ICSPC53359.2021.9689101.
[20] F. J. Adha, R. Ramli, M. H. Alkawaz, M. G. M. Johar, and A. I. Hajamydeen, “Assessment of conceptual framework for monitoring
poultry farm’s temperature and humidity,” in 2021 IEEE 11th International Conference on System Engineering and Technology,
ICSET 2021 - Proceedings, Nov. 2021, pp. 40–45, doi: 10.1109/ICSET53708.2021.9612437.
[21] M. Singh, R. Kumar, D. Tandon, P. Sood, and M. Sharma, “Artificial intelligence and IoT based monitoring of poultry health: a
review,” in 2020 IEEE International Conference on Communication, Networks and Satellite, Comnetsat 2020 - Proceedings, Dec.
2020, pp. 50–54, doi: 10.1109/Comnetsat50391.2020.9328930.
[22] K. Olejnik, E. Popiela, and S. Opaliński, “Emerging precision management methods in poultry sector,” Agriculture (Switzerland),
vol. 12, no. 5, p. 718, May 2022, doi: 10.3390/agriculture12050718.
[23] G. Franzo, M. Legnardi, G. Faustini, C. M. Tucciarone, and M. Cecchinato, “When everything becomes bigger: big data for big
poultry production,” Animals, vol. 13, no. 11, p. 1804, May 2023, doi: 10.3390/ani13111804.
[24] X. Li et al., “The genetic architecture of early body temperature and its correlation with salmonella pullorum resistance in three
chicken breeds,” Frontiers in Genetics, vol. 10, Jan. 2020, doi: 10.3389/fgene.2019.01287.
[25] I. Skomorucha and E. Sosnówka-Czajka, “The effect of adding herbal extracts to drinking water on body temperature, level of
thyroid hormones and H:L ratio in the blood of broiler chickens exposed to elevated ambient temperature,” Annals of Animal
Science, vol. 21, no. 4, pp. 1511–1522, Oct. 2021, doi: 10.2478/aoas-2021-0014.
[26] A. Goel, “Heat stress management in poultry,” Journal of Animal Physiology and Animal Nutrition, vol. 105, no. 6, pp. 1136–1145,
Nov. 2021, doi: 10.1111/jpn.13496.
[27] H. F. Saad, S. K. M. AL-hummod, S. M. Alsudany, and A. A. Jaffar, “The impact of high temperatures on the productive
performance (behavioral, physiological, and immunological) of poultry,” Texas Journal of Agriculture and Biological Sciences,
vol. 14, pp. 52–57, 2023.
[28] Z. Cai, J. Cui, H. Yuan, and M. Cheng, “Application and research progress of infrared thermography in temperature measurement
of livestock and poultry animals: a review,” Computers and Electronics in Agriculture, vol. 205, p. 107586, Feb. 2023,
doi: 10.1016/j.compag.2022.107586.
[29] J.-Y. Noh et al., “Thermal image scanning for the early detection of fever induced by highly pathogenic avian influenza virus
infection in chickens and ducks and its application in farms,” Frontiers in Veterinary Science, vol. 8, p. 616755, May 2021,
doi: 10.3389/fvets.2021.616755.
[30] C. H. Chuang, C. Y. Chiang, Y. C. Chen, C. Y. Lin, and Y. C. Tsai, “Goose surface temperature monitoring system based on deep
learning using visible and infrared thermal image integration,” IEEE Access, vol. 9, pp. 131203–131213, 2021,
doi: 10.1109/ACCESS.2021.3113509.

Poultry disease early detection methods using deep learning technology (Liu Yajie)
1722  ISSN: 2502-4752

[31] Z. Sun, M. Gao, M. Zhang, M. Lv, and G. Wang, “Research on recognition method of broiler overlapping sounds based on random
forest and confidence interval,” Computers and Electronics in Agriculture, vol. 209, p. 107801, Jun. 2023,
doi: 10.1016/j.compag.2023.107801.
[32] A. Banakar, M. Sadeghi, and A. Shushtari, “An intelligent device for diagnosing avian diseases: Newcastle, infectious bronchitis,
avian influenza,” Computers and Electronics in Agriculture, vol. 127, pp. 744–753, Sep. 2016, doi: 10.1016/j.compag.2016.08.006.
[33] M. M. D. Quintana, R. R. D. Infante, J. C. S. Torrano, and M. C. Pacis, “A hybrid solar powered chicken disease monitoring system
using decision tree models with visual and acoustic imagery,” in 2022 14th International Conference on Computer and Automation
Engineering (ICCAE), Mar. 2022, pp. 65–69, doi: 10.1109/ICCAE55086.2022.9762418.
[34] Z. Li et al., “Sex detection of chicks based on audio technology and deep learning methods,” Animals, vol. 12, no. 22, 2022,
doi: 10.3390/ani12223106.
[35] K. Cuan, T. Zhang, Z. Li, J. Huang, Y. Ding, and C. Fang, “Automatic Newcastle disease detection using sound technology and
deep learning method,” Computers and Electronics in Agriculture, vol. 194, p. 106740, Mar. 2022,
doi: 10.1016/j.compag.2022.106740.
[36] N. Jakovljevic, N. Maljkovic, D. Miskovic, P. Knezevic, and V. Delic, “A broiler stress detection system based on audio signal
processing,” in 27th Telecommunications Forum, TELFOR 2019, Nov. 2019, pp. 1–4, doi: 10.1109/TELFOR48224.2019.8971336.
[37] C. Wang, E. Benetos, S. Wang, and E. Versace, “Joint scattering for automatic chick call recognition,” in European Signal
Processing Conference, Aug. 2022, vol. 2022-Augus, pp. 195–199, doi: 10.23919/eusipco55093.2022.9909738.
[38] L. Carpentier, E. Vranken, D. Berckmans, J. Paeshuyse, and T. Norton, “Development of sound-based poultry health monitoring
tool for automated sneeze detection,” Computers and Electronics in Agriculture, vol. 162, pp. 573–581, Jul. 2019,
doi: 10.1016/j.compag.2019.05.013.
[39] J. Huang, W. Wang, and T. Zhang, “Method for detecting avian influenza disease of chickens based on sound analysis,” Biosystems
Engineering, vol. 180, pp. 16–24, Apr. 2019, doi: 10.1016/j.biosystemseng.2019.01.015.
[40] T. Tachibana, A. Nakatani, M. S. I. Khan, R. Makino, and M. A. Cline, “Behavioral and physiological responses to intraperitoneal
injection of zymosan in chicks,” Domestic Animal Endocrinology, vol. 72, p. 106471, Jul. 2020,
doi: 10.1016/j.domaniend.2020.106471.
[41] M. Z. Degu and G. L. Simegn, “Smartphone based detection and classification of poultry diseases from chicken fecal images using
deep learning techniques,” Smart Agricultural Technology, vol. 4, p. 100221, Aug. 2023, doi: 10.1016/j.atech.2023.100221.
[42] H. T. Gorji et al., “Combining deep learning and fluorescence imaging to automatically identify fecal contamination on meat
carcasses,” Scientific Reports, vol. 12, no. 1, p. 2392, Feb. 2022, doi: 10.1038/s41598-022-06379-1.
[43] J. Zhu and M. Zhou, “Online detection of abnormal chicken manure based on machine vision,” in 2021 ASABE Annual International
Virtual Meeting, July 12-16, 2021, 2021, vol. 1, pp. 595–601, doi: 10.13031/aim.202100188.
[44] H. O. Aworinde et al., “Poultry fecal imagery dataset for health status prediction: a case of South-West Nigeria,” Data in Brief,
vol. 50, p. 109517, Oct. 2023, doi: 10.1016/j.dib.2023.109517.
[45] H. Mbelwa, J. Mbelwa, and D. Machuve, “Deep convolutional neural network for chicken diseases detection,” International Journal
of Advanced Computer Science and Applications, vol. 12, no. 2, pp. 759–765, 2021, doi: 10.14569/IJACSA.2021.0120295.
[46] S. Suthagar, G. Mageshkumar, M. Ayyadurai, C. Snegha, M. Sureka, and S. Velmurugan, “Faecal image-based chicken disease
classification using deep learning techniques,” Lecture Notes in Networks and Systems, vol. 563, pp. 903–917, 2023,
doi: 10.1007/978-981-19-7402-1_64.
[47] W. Widyawati and W. Gunawan, “Early detection of sick chicken using artificial intelligence,” Teknika: Jurnal Sains dan Teknologi,
vol. 18, no. 2, p. 136, Nov. 2022, doi: 10.36055/tjst.v18i2.17337.
[48] J. Bao and Q. Xie, “Artificial intelligence in animal farming: A systematic literature review,” Journal of Cleaner Production,
vol. 331, p. 129956, Jan. 2022, doi: 10.1016/j.jclepro.2021.129956.
[49] L.-D. Quach, N. Pham-Quoc, D. C. Tran, and M. F. Hassan, “Identification of chicken diseases using VGGNet and ResNet models,”
in Lecture Notes of the Institute for Computer Sciences, Social-Informatics and Telecommunications Engineering, LNICST, vol.
334, 2020, pp. 259–269.
[50] A. H. Bin Hilmi, “Poultry pose estimation by detecting keypoints with deep convolutional network,” College of Engineering
Universiti Tenaga Nasional, 2022.
[51] C. Fang, H. Zheng, J. Yang, H. Deng, and T. Zhang, “Study on poultry pose estimation based on multi-parts detection,” Animals,
vol. 12, no. 10, p. 1322, May 2022, doi: 10.3390/ani12101322.
[52] C. Fang, T. Zhang, H. Zheng, J. Huang, and K. Cuan, “Pose estimation and behavior classification of broiler chickens based on
deep neural networks,” Computers and Electronics in Agriculture, vol. 180, p. 105863, Jan. 2021, doi: 10.1016/j.compag.2020.105863.
[53] A. Nasiri, J. Yoder, Y. Zhao, S. Hawkins, M. Prado, and H. Gan, “Pose estimation-based lameness recognition in broiler using
CNN-LSTM network,” Computers and Electronics in Agriculture, vol. 197, p. 106931, Jun. 2022, doi: 10.1016/j.compag.2022.106931.
[54] B. X. Xie and C. L. Chang, “Behavior recognition of a broiler chicken using long short-term memory with convolution neural
networks,” in 2022 International Automatic Control Conference, CACS 2022, Nov. 2022, pp. 1–5,
doi: 10.1109/CACS55319.2022.9969848.
[55] H. Zhang and C. Chen, “Design of sick chicken automatic detection system based on improved residual network,” in Proceedings
of 2020 IEEE 4th Information Technology, Networking, Electronic and Automation Control Conference, ITNEC 2020, Jun. 2020,
pp. 2480–2485, doi: 10.1109/ITNEC48623.2020.9084666.
[56] N. Volkmann et al., “Keypoint detection for injury identification during Turkey Husbandry using neural networks,” Sensors,
vol. 22, no. 14, p. 5188, Jul. 2022, doi: 10.3390/s22145188.
[57] M. A. A. A. Bakar, P. J. Ker, S. G. H. Tang, H. J. Lee, and B. S. Zainal, “Classification of unhealthy chicken based on chromaticity
of the comb,” in 2022 IEEE International Conference on Computing, ICOCO 2022, Nov. 2022, pp. 1–5,
doi: 10.1109/ICOCO56118.2022.10031812.
[58] M. K. Gourisaria, A. Arora, S. Bilgaiyan, and M. Sahni, “Chicken disease multiclass classification using deep learning,” in Lecture
Notes in Networks and Systems, vol. 614 LNNS, 2023, pp. 225–238.
[59] X. Yang, R. Bist, S. Subedi, Z. Wu, T. Liu, and L. Chai, “An automatic classifier for monitoring applied behaviors of cage-free
laying hens with deep learning,” Engineering Applications of Artificial Intelligence, vol. 123, p. 106377, Aug. 2023,
doi: 10.1016/j.engappai.2023.106377.
[60] H. Kim, W. D. Lee, H. K. Jang, M. Kang, and H. K. Kang, “The potential of non-movement behavior observation method for
detection of sick broiler chickens,” Journal of Animal Science and Technology, vol. 65, no. 2, pp. 441–458, Dec. 2023,
doi: 10.5187/jast.2022.e105.

Indonesian J Elec Eng & Comp Sci, Vol. 32, No. 3, December 2023: 1712-1723
Indonesian J Elec Eng & Comp Sci ISSN: 2502-4752  1723

BIOGRAPHIES OF AUTHORS

Liu Yajie is pursuing a Ph.D. degree in Computer Science at Management and

Science University, Malaysia. He earned his B.S. degree in Software Engineering from
Henan University of Economics and Law, China, in 2018. Following that, he obtained an
M.Tech. degree in Computer Science from SRM Institute of Science and Technology, India,
in 2020. His research interests include internet of things (IoT), image processing, and pattern
recognition. He can be contacted at email: [email protected].

Md Gapar Md Johar is senior vice President Research, Innovation, Technology

and System of Management and Science University, Malaysia. He is a Professor in software
engineering. Ph.D. in computer science, M.Sc. in data engineering, and B.Sc. (Hons) in
computer science and is a certified e-commerce consultant are among his qualifications.
Working and teaching in several organizations for more than 40 years has given him a broad
range of expertise. His research interests include data mining, learning content management
system, blended assessment system, knowledge management system, RFID, e-commerce,
character recognition, image processing, and healthcare management system. He can be
contacted at email: [email protected].

Asif Iqbal Hajamydeen is Associate Professor at Faculty of Information

Sciences and Engineering, Management and Science University, Shah Alam, Selangor,
Malaysia. He holds a Ph.D. degree in Security in Computing in 2014. He earned his M.Sc.,
Computer Science, Bharathidasan University in 1998. His research areas are data mining, and
anomaly detection. He can be contacted at email: [email protected].

Poultry disease early detection methods using deep learning technology (Liu Yajie)