Aqeel Raza
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Aqeel Raza has a total of 26 co-authors. Here are the top co-authors by number of shared publications:
- Chaidate Inchaisri
Center of Excellence in Data Innovation for Sustainable Livestock Production (CEDISLP), Department of Veterinary Medicine, Faculty of Veterinary Science, Chulalongkorn University, Bangkok 10330, Thailand
7 shared publications - Kumail Abbas
International Graduate Program of Veterinary Science and Technology Faculty of Veterinary Science Chulalongkorn University 10440 Bangkok Thailand
6 shared publications - Henk Hogeveen
Wageningen University & Research
3 shared publications - Theerawat Swangchan-Uthai
Department of Obstetrics, Gynaecology and Reproduction, Faculty of Veterinary Science, Chulalongkorn University, Bangkok 10330, Thailand; Center of Excellence in Animal Fertility Chulalongkorn University (CU-AF), Chulalongkorn University, Bangkok 10330, Thailand
3 shared publications - Ali Alameer
School of Science, Engineering and Environment, University of Salford, Salford, UK
2 shared publications - Andrew W. Dowsey
Bristol Veterinary School, University of Bristol, Bristol BS40 5DU, UK
2 shared publications
Latest Publications
- Open Access
- 6 October 2026
- Published by Elsevier in Smart Agricultural Technology
- Vol. 15
A Hierarchical Computer Vision Framework for Continuous Dairy Cow Behavior Monitoring in an Obstructed Free-Stall Barns
- 6 October 2026
- Published by Elsevier in Smart Agricultural Technology
- Vol. 15
Open AccessA Hierarchical Computer Vision Framework for Continuous Dairy Cow Behavior Monitoring in an Obstructed Free-Stall Barns
- Open Access
- 23 September 2026
- Published by Elsevier in SSRN Electronic Journal
Multi-Camera Computer Vision Reveals Cow-Specific Feeding Responses and Longitudinal Behavioural Patterns in Transition Dairy Cows
Automated monitoring of dairy cow behaviour has advanced rapidly, but most camera-based systems still provide group-level summaries rather than biologically interpretable profiles of individual animals over time. This limits their value during the transition period, when changes in feeding, resting and activity patterns may provide early evidence of compromised welfare. In this study we developed a multi-camera computer-vision framework to convert routine CCTV footage from a group-housed transition pen into individual longitudinal behavioural profiles. Across 12 consecutive observation days, the system tracked cows, classified feeding, drinking, standing and resting behaviour, and summarised each cow’s response around morning feed delivery. The final smoothed MaxViT-T classifier achieved 91.2% accuracy and a macro F1-score of 0.866 in the temporal validation against manually labelled video. The stable 9-cow cohort contributed 99 cow-day records across 11 days. Median feeding proportion increased from a median of 0.0% before feed delivery to 87.4% during the Feeding window, while resting and standing declined. Feeding and resting showed a weak positive association (r = 0.215, p = 0.033), while clear between-cow heterogeneity was evident in longitudinal profiles, but repeatability was low across all eight outcomes (R = 0.000–0.065). The system also distinguished a shared cohort-level response on the earliest feed-delivery day from individual-specific deviations. These findings demonstrate that CCTV-based monitoring can support biologically meaningful individual behavioural profiling in transition dairy cows. Further validation against clinical health outcomes, together with testing across farms, is required before production use for decision-support.
- 23 September 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessMulti-Camera Computer Vision Reveals Cow-Specific Feeding Responses and Longitudinal Behavioural Patterns in Transition Dairy Cows
Automated monitoring of dairy cow behaviour has advanced rapidly, but most camera-based systems still provide group-level summaries rather than biologically interpretable profiles of individual animals over time. This limits their value during the transition period, when changes in feeding, resting and activity patterns may provide early evidence of compromised welfare. In this study we developed a multi-camera computer-vision framework to convert routine CCTV footage from a group-housed transition pen into individual longitudinal behavioural profiles. Across 12 consecutive observation days, the system tracked cows, classified feeding, drinking, standing and resting behaviour, and summarised each cow’s response around morning feed delivery. The final smoothed MaxViT-T classifier achieved 91.2% accuracy and a macro F1-score of 0.866 in the temporal validation against manually labelled video. The stable 9-cow cohort contributed 99 cow-day records across 11 days. Median feeding proportion increased from a median of 0.0% before feed delivery to 87.4% during the Feeding window, while resting and standing declined. Feeding and resting showed a weak positive association (r = 0.215, p = 0.033), while clear between-cow heterogeneity was evident in longitudinal profiles, but repeatability was low across all eight outcomes (R = 0.000–0.065). The system also distinguished a shared cohort-level response on the earliest feed-delivery day from individual-specific deviations. These findings demonstrate that CCTV-based monitoring can support biologically meaningful individual behavioural profiling in transition dairy cows. Further validation against clinical health outcomes, together with testing across farms, is required before production use for decision-support.
- Open Access
- 27 July 2026
- Published by Elsevier in SSRN Electronic Journal
A Hierarchical Computer Vision Framework for Continuous Dairy Cow Behavior Monitoring in an Obstructed Free-Stall Barns
Continuous monitoring of dairy cows in free-stall barns remains challenging because of occlusion, variable illumination, and identity switches. This study developed a hierarchical computer vision framework integrating behavior-specific object detection, detector-assisted SAM2 segmentation, and multi-object tracking for CCTV-based monitoring. Video data were collected from three CCTV cameras in free-stall barns, covering two obstructed zone-specific views, and one wide-angle facility view. YOLOv11m and RT-DETR-L were compared across five behaviors (drinking, feeding, resting, standing, and walking) using a 214,187-image annotated detection dataset, while 645 daytime and 283 nighttime images evaluated detector-assisted SAM2 segmentation. Tracking performance was assessed using ByteTrack and SAM2-assisted tracking under variable illumination and camera-view conditions. RT-DETR-L achieved a higher F1-score than YOLOv11m (66.0% vs. 63.1%) despite identical mAP50 (65.4%) and mAP50-95 (50.9%). Feeding, resting, and standing were detected more reliably than walking and drinking, with drinking showing the lowest AP50 (2.0%). For segmentation, RT-DETR-L-SAM2 achieved higher and more stable mean IoU (60.5% and 59.4%) than YOLOv11m-SAM2 (44.9% and 50.5%) across daytime and nighttime conditions. RT-DETR-L combined with ByteTrack achieved the highest MOTA (53.0%) and IDF1 (74.6%). SAM2-assisted tracking substantially reduced identity switches, from 50 to 8 for YOLOv11m and from 75 to 9 for RT-DETR-L. This improvement did not consistently raise aggregate MOTA or IDF1 but markedly enhance identity continuity, revealing a trade-off between tracking accuracy and long-term identity preservation. These findings demonstrate that integrating detection, SAM2-based segmentation, and tracking provides a robust framework for continuous CCTV-based dairy cow monitoring, particularly maintaining individual identities in complex, dynamic free-stall barns.
- 27 July 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessA Hierarchical Computer Vision Framework for Continuous Dairy Cow Behavior Monitoring in an Obstructed Free-Stall Barns
Continuous monitoring of dairy cows in free-stall barns remains challenging because of occlusion, variable illumination, and identity switches. This study developed a hierarchical computer vision framework integrating behavior-specific object detection, detector-assisted SAM2 segmentation, and multi-object tracking for CCTV-based monitoring. Video data were collected from three CCTV cameras in free-stall barns, covering two obstructed zone-specific views, and one wide-angle facility view. YOLOv11m and RT-DETR-L were compared across five behaviors (drinking, feeding, resting, standing, and walking) using a 214,187-image annotated detection dataset, while 645 daytime and 283 nighttime images evaluated detector-assisted SAM2 segmentation. Tracking performance was assessed using ByteTrack and SAM2-assisted tracking under variable illumination and camera-view conditions. RT-DETR-L achieved a higher F1-score than YOLOv11m (66.0% vs. 63.1%) despite identical mAP50 (65.4%) and mAP50-95 (50.9%). Feeding, resting, and standing were detected more reliably than walking and drinking, with drinking showing the lowest AP50 (2.0%). For segmentation, RT-DETR-L-SAM2 achieved higher and more stable mean IoU (60.5% and 59.4%) than YOLOv11m-SAM2 (44.9% and 50.5%) across daytime and nighttime conditions. RT-DETR-L combined with ByteTrack achieved the highest MOTA (53.0%) and IDF1 (74.6%). SAM2-assisted tracking substantially reduced identity switches, from 50 to 8 for YOLOv11m and from 75 to 9 for RT-DETR-L. This improvement did not consistently raise aggregate MOTA or IDF1 but markedly enhance identity continuity, revealing a trade-off between tracking accuracy and long-term identity preservation. These findings demonstrate that integrating detection, SAM2-based segmentation, and tracking provides a robust framework for continuous CCTV-based dairy cow monitoring, particularly maintaining individual identities in complex, dynamic free-stall barns.
- Open Access
- 9 October 2025
- Published by Elsevier in Smart Agricultural Technology
- Vol. 12
Vision transformer-based multi-camera multi-object tracking framework for dairy cow monitoring
- 9 October 2025
- Published by Elsevier in Smart Agricultural Technology
- Vol. 12
Open AccessVision transformer-based multi-camera multi-object tracking framework for dairy cow monitoring
- Open Access
- 13 May 2025
- Published by Veterinary World in Veterinary World
- Vol. 18 (5) , 1109-1126
Impacts of early postpartum behavioral patterns on the fertility and milk production of tropical dairy cows
Background and Aim: Early postpartum behavioral patterns are pivotal indicators of dairy cow health, reproductive success, and lactation performance, particularly under the environmental stressors of tropical climates. This study aimed to investigate how these behavioral patterns, as captured by smart biosensor data, influence reproductive outcomes, and milk yield in Holstein Friesian cows, with specific emphasis on parity differences and behavioral clustering. Materials and Methods: A total of 227 Holstein Friesian cows, categorized by parity (primiparous vs. multiparous), were monitored using AfiTag-II accelerometers from 3 days prepartum to 30 days postpartum. Behavioral variables – activity, rest time, rest per bout, and restlessness ratio – were subjected to K-means clustering to identify distinct behavioral profiles. Reproductive performance was analyzed using Cox proportional hazard models, while lactation dynamics were modeled using the Wood function to estimate peak yield, peak time, and persistency. Results: Three distinct behavioral clusters were identified. Primiparous cows in Cluster 1 showed the highest early postpartum activity (~300 min/day at 5 days in milk [DIM]) and restlessness ratios, while multiparous cows exhibited more stable behavioral profiles. Cox regression suggested that cows in Cluster 0 had a higher, although non-significant, likelihood of estrus onset at 40 DIM (Hazard ratio = 1.44, p = 0.09). Lactation modeling revealed that multiparous cows in Cluster 0 attained the highest cumulative milk yield (4896.6 ± 252.1 kg at 305 DIM), while the single cow in Cluster 2 exhibited an atypical lactation curve with a delayed peak and reduced persistency. Conclusion: Postpartum behavioral clustering reveals parity-specific lactation and reproductive trajectories in tropical dairy cows. Higher activity and restlessness ratios may delay estrus and compromise milk yield, underscoring the potential of behavioral monitoring for targeted reproductive and nutritional management. Integration of sensor-based clustering with routine herd monitoring may support early identification of cows at risk of suboptimal performance, improving reproductive efficiency and milk production in tropical dairy systems. Keywords: K-means clustering, milk yield, parity, postpartum behavior, reproductive performance, smart sensors, tropical dairy cattle.
- 13 May 2025
- Published by Veterinary World in Veterinary World
- Vol. 18 (5) , 1109-1126
Open AccessImpacts of early postpartum behavioral patterns on the fertility and milk production of tropical dairy cows
Background and Aim: Early postpartum behavioral patterns are pivotal indicators of dairy cow health, reproductive success, and lactation performance, particularly under the environmental stressors of tropical climates. This study aimed to investigate how these behavioral patterns, as captured by smart biosensor data, influence reproductive outcomes, and milk yield in Holstein Friesian cows, with specific emphasis on parity differences and behavioral clustering. Materials and Methods: A total of 227 Holstein Friesian cows, categorized by parity (primiparous vs. multiparous), were monitored using AfiTag-II accelerometers from 3 days prepartum to 30 days postpartum. Behavioral variables – activity, rest time, rest per bout, and restlessness ratio – were subjected to K-means clustering to identify distinct behavioral profiles. Reproductive performance was analyzed using Cox proportional hazard models, while lactation dynamics were modeled using the Wood function to estimate peak yield, peak time, and persistency. Results: Three distinct behavioral clusters were identified. Primiparous cows in Cluster 1 showed the highest early postpartum activity (~300 min/day at 5 days in milk [DIM]) and restlessness ratios, while multiparous cows exhibited more stable behavioral profiles. Cox regression suggested that cows in Cluster 0 had a higher, although non-significant, likelihood of estrus onset at 40 DIM (Hazard ratio = 1.44, p = 0.09). Lactation modeling revealed that multiparous cows in Cluster 0 attained the highest cumulative milk yield (4896.6 ± 252.1 kg at 305 DIM), while the single cow in Cluster 2 exhibited an atypical lactation curve with a delayed peak and reduced persistency. Conclusion: Postpartum behavioral clustering reveals parity-specific lactation and reproductive trajectories in tropical dairy cows. Higher activity and restlessness ratios may delay estrus and compromise milk yield, underscoring the potential of behavioral monitoring for targeted reproductive and nutritional management. Integration of sensor-based clustering with routine herd monitoring may support early identification of cows at risk of suboptimal performance, improving reproductive efficiency and milk production in tropical dairy systems. Keywords: K-means clustering, milk yield, parity, postpartum behavior, reproductive performance, smart sensors, tropical dairy cattle.
- Open Access
Behavioral Adaptations in Tropical Dairy Cows: Insights into Calving Day Predictions
This study examined changes in the activity patterns of tropical dairy cows during the transition period to assess their potential for predicting calving days. This study used the AfiTag-II biosensor to monitor activity, rest time, rest per bout, and restlessness ratio in 298 prepartum and 347 postpartum Holstein Friesian cows across three lactation groups (1, 2, and ≥3). The data were analyzed using generalized linear mixed models in SPSS, and five machine learning models, including random forest, decision tree, gradient boosting, Naïve Bayes, and neural networks, were used to predict the calving day, with their performance evaluated via ROC curves and AUC metrics. For all lactations, activity levels peak on the calving day, followed by a gradual return to prepartum levels within two weeks. First-lactation cows displayed the shortest rest duration, with a prepartum rest time of 568.8 ± 5.4 (mean ± SE), which is significantly lower than higher-lactation animals. The random forest and gradient boosting displayed an effective performance, achieving AUCs of 85% and 83%, respectively. These results indicate that temporal changes in activity behavior have the potential to be a useful indicator for calving day prediction, particularly in tropical climates where seasonal variations can obscure traditional prepartum indicators.
Open AccessBehavioral Adaptations in Tropical Dairy Cows: Insights into Calving Day Predictions
This study examined changes in the activity patterns of tropical dairy cows during the transition period to assess their potential for predicting calving days. This study used the AfiTag-II biosensor to monitor activity, rest time, rest per bout, and restlessness ratio in 298 prepartum and 347 postpartum Holstein Friesian cows across three lactation groups (1, 2, and ≥3). The data were analyzed using generalized linear mixed models in SPSS, and five machine learning models, including random forest, decision tree, gradient boosting, Naïve Bayes, and neural networks, were used to predict the calving day, with their performance evaluated via ROC curves and AUC metrics. For all lactations, activity levels peak on the calving day, followed by a gradual return to prepartum levels within two weeks. First-lactation cows displayed the shortest rest duration, with a prepartum rest time of 568.8 ± 5.4 (mean ± SE), which is significantly lower than higher-lactation animals. The random forest and gradient boosting displayed an effective performance, achieving AUCs of 85% and 83%, respectively. These results indicate that temporal changes in activity behavior have the potential to be a useful indicator for calving day prediction, particularly in tropical climates where seasonal variations can obscure traditional prepartum indicators.
- Open Access
- 8 June 2024
- Published by Veterinary World in Veterinary World
- Vol. 17 (6) , 1251-1258
Smart sensors in Thai dairy reproduction: A case study
Background and Aim: Movement activity sensors are known for their potential to boost the reproductive performance of dairy cows. This study evaluated the effectiveness of these sensors on three Thai dairy farms (MK, NF, and CC), each using different sensor brands. We focused on reproductive performance at these farms and expanded our evaluation to include farmer satisfaction with sensor technology on five farms (MK, NF, CC, AP, and IP), allowing for a thorough analysis of both operational outcomes and user feedback. Materials and Methods: A total of 298 lactation records and interviewing five experienced farm owners with over a year of sensor usage were our research methods. To measure the effect on the first service timing and post-parturition pregnancy rates, Cox regression models were utilized for sensor usage. Results: Biosensors’ implementation enhanced data precision while quickening the first service within 100 days and pregnancy within 200 days. The MK and NF farms showed significant progress. Within 100 and 200 days post-implementation, the overall improvement was 30%–34% in the first service rate and 39%–67% in the conception rate across all assessed farms. Farmers acknowledged improved reproductive performance from the sensors, overcoming language barriers. Conclusion: The study highlighted the advantages of using movement activity sensors in enhancing both cattle reproductive success and farmers’ satisfaction on Thai dairy farms. These sensors led to more accurate management decisions, increasing overall farm productivity. Keywords: dairy cattle, movement activity sensors, reproductive performance.
- 8 June 2024
- Published by Veterinary World in Veterinary World
- Vol. 17 (6) , 1251-1258
Open AccessSmart sensors in Thai dairy reproduction: A case study
Background and Aim: Movement activity sensors are known for their potential to boost the reproductive performance of dairy cows. This study evaluated the effectiveness of these sensors on three Thai dairy farms (MK, NF, and CC), each using different sensor brands. We focused on reproductive performance at these farms and expanded our evaluation to include farmer satisfaction with sensor technology on five farms (MK, NF, CC, AP, and IP), allowing for a thorough analysis of both operational outcomes and user feedback. Materials and Methods: A total of 298 lactation records and interviewing five experienced farm owners with over a year of sensor usage were our research methods. To measure the effect on the first service timing and post-parturition pregnancy rates, Cox regression models were utilized for sensor usage. Results: Biosensors’ implementation enhanced data precision while quickening the first service within 100 days and pregnancy within 200 days. The MK and NF farms showed significant progress. Within 100 and 200 days post-implementation, the overall improvement was 30%–34% in the first service rate and 39%–67% in the conception rate across all assessed farms. Farmers acknowledged improved reproductive performance from the sensors, overcoming language barriers. Conclusion: The study highlighted the advantages of using movement activity sensors in enhancing both cattle reproductive success and farmers’ satisfaction on Thai dairy farms. These sensors led to more accurate management decisions, increasing overall farm productivity. Keywords: dairy cattle, movement activity sensors, reproductive performance.
- Open Access
- 21 September 2022
- Published by MDPI AG in Biomolecules
- Vol. 12 (10) , 1342
Ractopamine at the Center of Decades-Long Scientific and Legal Disputes: A Lesson on Benefits, Safety Issues, and Conflicts
Ractopamine (RAC) is a synthetic phenethanolamine, β–adrenergic agonist used as a feed additive to develop leanness and increase feed conversion efficiency in different farm animals. While RAC has been authorized as a feed additive for pigs and cattle in a limited number of countries, a great majority of jurisdictions, including the European Union (EU), China, Russia, and Taiwan, have banned its use on safety grounds. RAC has been under long scientific and political discussion as a controversial antibiotic as a feed additive. Here, we will present significant information on RAC regarding its application, detection methods, conflicts, and legal divisions that play a major role in controversial deadlock and why this issue warrants the attention of scientists, agriculturists, environmentalists, and health advocates. In this review, we highlight the potential toxicities of RAC on aquatic animals to emphasize scientific evidence and reports on the potentially harmful effects of RAC on the aquatic environment and human health.
- 21 September 2022
- Published by MDPI AG in Biomolecules
- Vol. 12 (10) , 1342
Open AccessRactopamine at the Center of Decades-Long Scientific and Legal Disputes: A Lesson on Benefits, Safety Issues, and Conflicts
Ractopamine (RAC) is a synthetic phenethanolamine, β–adrenergic agonist used as a feed additive to develop leanness and increase feed conversion efficiency in different farm animals. While RAC has been authorized as a feed additive for pigs and cattle in a limited number of countries, a great majority of jurisdictions, including the European Union (EU), China, Russia, and Taiwan, have banned its use on safety grounds. RAC has been under long scientific and political discussion as a controversial antibiotic as a feed additive. Here, we will present significant information on RAC regarding its application, detection methods, conflicts, and legal divisions that play a major role in controversial deadlock and why this issue warrants the attention of scientists, agriculturists, environmentalists, and health advocates. In this review, we highlight the potential toxicities of RAC on aquatic animals to emphasize scientific evidence and reports on the potentially harmful effects of RAC on the aquatic environment and human health.