Object-Assisted Question Featurization and Multi-CNN Image Feature Fusion for Visual Question Answering

Visual question answering (VQA) demands a meticulous and concurrent proficiency in image interpretation natural language understanding to correctly answer the about an image. The existing VQA solutions either focus only on improving joint multi-modal embedding or fine-tuning of visual through attention. This research, contrast current trend, investigates feasibility object-assisted strategy titled semantic object ranking (SOR) framework for VQA. proposed system refines representation with help detected objects. For multi-CNN representation, employs canonical correlation analysis (CCA). suggested model is assessed using accuracy WUPS measures DAQUAR dataset. On dataset, analytical outcomes reveal that presented outperforms prior state-of-the-art by significant factor. In addition quantitative analysis, proper illustrations are supplied observe reasons performance improvement.