Analyzing Data Modalities for Cattle Weight Estimation Using Deep Learning Models.

We investigate the impact of different data modalities for cattle weight estimation. For this purpose, we collect and present our own cattle dataset representing the data modalities: RGB, depth, combined RGB and depth, segmentation, and combined segmentation and depth information. We explore a recent vision-transformer-based zero-shot model proposed by Meta AI Research for producing the segmentation data modality and for extracting the cattle-only region from the images.

Use of immobilized cryopreserved bovine semen in a blind artificial insemination trial.

To make timing of artificial insemination (AI) relative to ovulation less critical, methods for prolonging shelf life of spermatozoa in vivo after AI have been attempted to be developed. Encapsulation of sperm cells is a documented technology, and recently, a technology in which sperm cells are embedded in alginate gel has been introduced and commercialized. In this study, standard processed semen with the Biladyl extender (control) was compared with semen processed by sperm immobilization technology developed by SpermVital AS in a blind field trial.