Author ORCID Identifier:

https://orcid.org/0009-0003-9472-1182

Date of Graduation

8-2025

Document Type

Thesis

Degree Name

Master of Science in Computer Science (MS)

Degree Level

Graduate

Department

Computer Science & Computer Engineering

Advisor/Mentor

Le, Ngan

Committee Member

Gauch, John

Second Committee Member

Zhang, Lu

Keywords

Chick Sexing; Computer Vision; Deep Learning; Facial; Poultry; Vent (cloaca)

Abstract

This thesis presents two complementary approaches to chick sexing, a critical task in poultry production that demands accurate and early gender identification. By investigating both facial and vent-based modalities, we present two complementary methods that aim to improve the efficiency, scalability, and ethical standards of gender classification in day-old chicks through the use of computer vision and deep learning. The first approach draws inspiration from human facial gender recognition to introduce facial chick sexing, a minimally invasive technique that eliminates the need for expert knowledge. This system encompasses a complete pipeline that includes image acquisition, facial detection and alignment, keypoint localization, and final classification. Two facial cropping strategies are evaluated—Cropped Full Face and Cropped Middle Face—both of which retain critical visual features. This method demonstrates promising accuracy while offering a scalable and welfare-conscious alternative to traditional practices. In addition, we propose VentVision an automated system for vent-based chick sexing, automating the process through a structured pipeline comprising multi-modality video capture, temporal segment selection, vent region detection, and gender classification. The proposed system incorporates both RGB and infrared imaging to enhance feature extraction and improve model reliability. Experiments show that the combined RGB-IR modality achieves expert-level accuracy, reinforcing the method's potential for deployment in real-world environments. Together, these methods offer two distinct yet complementary pathways toward modernizing chick sexing: one grounded in facial biometrics and another leveraging enhanced vent analysis. Extensive quantitative and qualitative experiments confirm the effectiveness of each approach. In conclusion, this thesis contributes a novel framework for facial chick sexing and an automated implementation of traditional vent sexing, both designed to balance precision, efficiency, and ethical considerations. By reducing dependence on specialized expertise and advancing automation in chick sexing, we pave the way for scalable and humane solutions in poultry production.

Download

Share

COinS