Vol. 3 No. 7 (2026)

Issue 7 of the International Journal of Emerging Technologies and Advanced Applications features work on applying computer vision to practical aquaculture problems. Feliciano and Damian address a persistent difficulty in hatchery management: counting eels, whose elongated bodies and transparency at the glass eel stage make them unusually prone to occlusion and duplicate detection. Rather than redesigning the detection network, the authors optimize confidence and Intersection over Union thresholds during post-processing, raising mean counting accuracy from 80.95% to 92.24% on an independent evaluation set. The approach is lightweight, deployable via ONNX, and points toward a broader principle: careful parameter tuning can substitute for architectural complexity.