https://www.nxn.ai/Utilize Generative AI

Exchange student to NYU CAS (GPA: 4.0)

Domain: Computer vision - image detection/segmentationKeywords: Transformer, ViT, DETR- Qualitative and quantitative analysis of each model after applying multiple transformer-based encoders (ViT, XCiT, DEiT, Swin Transformer, Segformer, ViT Adapter) and decoders (Segmenter, Mask2Former) for segmentation.- Change the framework of the cell detection task from a heatmap-based framework to a regression-based framework utilizing DETR.- Understand the different detr-based models (DETR, Conditional DETR, Deformable DETR, DAB-DETR*, DN-DETR, DINO, Mask DINO) and analyze the advantages and disadvantages of each model for downstream tasks.- Develop a visualizer for qualitative analysis for the new cell detection framework.- Propose a new transformer-based architecture that can synthesize the context between Multi-FoV (Field of View) images to improve the performance of cell detection.