Source author record

Harish Katti

Harish Katti appears in the imported research catalog. Authorship, coauthor and topic links are available while profile ownership is still unclaimed.

ResearcherUnclaimed source record

Computer Vision Human-Computer Interaction Neurons and Cognition

Catalog footprint

What is connected

3works

3topics

4close collaborators

Actions

Connect this record

Open graph Browse works

Inspect adjacent papers, topics, institutions and collaborators without losing the researcher page.

Building this map preview

BZPEER is loading the nearby papers, people, topics and institutions for this page.

preprint2022arXiv

Affective Computational Advertising Based on Perceptual Metrics

We present \textbf{ACAD}, an \textbf{a}ffective \textbf{c}omputational \textbf{ad}vertising framework expressly derived from perceptual metrics. Different from advertising methods which either ignore the emotional nature of (most) programs and ads, or are based on axiomatic rules, the ACAD formulation incorporates findings from a user study examining the effect of within-program ad placements on ad perception. A linear program formulation seeking to achieve (a) \emph{genuine} ad assessments and (b) \emph{maximal} ad recall is then proposed. Effectiveness of the ACAD framework is confirmed via a validational user study, where ACAD-induced ad placements are found to be optimal with respect to objectives (a) and (b) against competing approaches.

preprint2020arXiv

Human peripheral blur is optimal for object recognition

Our vision is sharpest at the center of our gaze and becomes progressively blurry into the periphery. It is widely believed that this high foveal resolution evolved at the expense of peripheral acuity. But what if this sampling scheme is actually optimal for object recognition? To test this hypothesis, we trained deep neural networks on 'foveated' images with high resolution near objects and increasingly sparse sampling into the periphery. Neural networks trained using a blur profile matching the human eye yielded the best performance compared to shallower and steeper blur profiles. Even in humans, categorization accuracy deteriorated only for steeper blur profiles. Thus, our blurry peripheral vision may have evolved to optimize object recognition rather than merely due to wiring constraints.

preprint2020arXiv

Predicting population neural activity in the Algonauts challenge using end-to-end trained Siamese networks and group convolutions

The Algonauts challenge is about predicting the object representations in the form of Representational Dissimilarity Matrices (RDMS) derived from visual brain regions. We used a customized deep learning model using the concept of Siamese networks and group convolutions to predict neural distances corresponding to a pair of images. Training data was best explained by distances computed over the last layer.