Papers
5
Total Citations
712
H-Index
4
About
Ali Borji is a leading researcher in visual attention, saliency modeling, and deep learning for object recognition. His most influential work, a 2012 comparative study on human-model agreement in visual saliency, has garnered over 620 citations, establishing a benchmark for evaluating how well computational models predict where humans look in images. Borji’s research bridges biological vision and machine learning, exploring both bottom-up saliency (image-driven attention) and top-down task influences. He also created the iLab-20M dataset, a large-scale controlled object dataset with over 20 million images, designed to systematically test deep neural networks’ tolerance to variations in translation, scale, pose, and illumination. This resource has been pivotal for understanding CNN limitations and guiding robust model design. Beyond saliency, Borji has contributed to gesture recognition for interactive robotics and sequential attention control, applying saliency maps to real-time hand tracking and robotic perception. His work has shaped how researchers evaluate and improve visual attention models, making him a key figure in both computational neuroscience and computer vision.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3
- 4
- 5