From Semantic To Instance: A Semi-Self-Supervised Learning Approach

ArXi:2506.16563v2 Announce Type: replace-cross Instance segmentation is essential for applications such as automated monitoring of plant health, growth, and yield. However, extensive effort is required to create large-scale datasets with pixel-level annotations of each object instance for developing instance segmentation models that restrict the use of deep learning in these areas. This challenge is significant in images with densely packed, self-occluded objects, which are common in agriculture.