The field of computer vision is rapidly advancing, with a focus on developing innovative methods for object detection and analysis in various domains, including microscopy and environmental monitoring. Recent developments have centered around improving the efficiency and accuracy of object detection algorithms, particularly in cases where annotated data is scarce or difficult to obtain. Researchers are exploring the use of weakly supervised learning, self-supervised learning, and transfer learning to adapt models to new tasks and domains. Additionally, there is a growing interest in applying computer vision techniques to real-world problems, such as microplastic detection and analysis, and vehicle search and re-identification. Noteworthy papers in this area include: Weakly Supervised Virus Capsid Detection with Image-Level Annotations in Electron Microscopy Images, which proposes a domain-specific weakly supervised object detection algorithm. GECO: Geometrically Consistent Embedding with Lightspeed Inference, which introduces a training framework based on optimal transport to produce geometrically coherent features. Zero-shot Shape Classification of Nanoparticles in SEM Images using Vision Foundation Models, which achieves high-precision shape classification without extensive labeled datasets. ParticleSAM: Small Particle Segmentation for Material Quality Monitoring in Recycling Processes, which adapts the segmentation foundation model to images with small and dense objects. 4D-PreNet: A Unified Preprocessing Framework for 4D-STEM Data Analysis, which integrates attention-enhanced U-Net and ResNet architectures for denoising, center correction, and elliptical distortion calibration. CLIPVehicle: A Unified Framework for Vision-based Vehicle Search, which proposes a dual-granularity semantic-region alignment module and a multi-level vehicle identification learning strategy. Segment Any Vehicle: Semantic and Visual Context Driven SAM and A Benchmark, which introduces a novel framework comprising a SAM-based encoder-decoder, a vehicle part knowledge graph, and a context sample retrieval encoding module. Segmenting the Complex and Irregular in Two-Phase Flows: A Real-World Empirical Study with SAM2, which demonstrates the effectiveness of a fine-tuned Segment Anything Model in segmenting highly non-convex, irregular bubble structures.