Sunil Jaiswal

I am working as a computer vision researcher at K|Lens GmbH and also leading the computer vision-AI R&D team focussing on real-world computer vision problems in bringing ideas into reality from research prototypes to a product. My research interest includes:

I am fortunate enough to work with some great minds:

This work shows how varied random degradations can contribute to learning an effective VSR model, especially for real-world video artifacts.

In this work, we generate a high-resolution synthetic depth dataset (HRSD) which contains 100,000 color images and corresponding dense ground truth depth maps.

In this work, We introduce a novel multi-view dataset with semi-automatic ground-truth data, which results in significant labeling resource savings.

In contrast to the naive method of estimating depths from images, a more sophisticated approach uses temporal information, thereby eliminating flickering and geometrical inconsistencies. We propose a consistent method for dense video depth estimation; however, unlike the existing monocular methods, ours relates to stereo videos.

In this work, we propose a new and efficient quality assessment algorithm based upon the variation in the depth of 3D synthesized and reference views.

The applicability to real-world images is limited and expectations are often disappointed when comparing to the performance achieved on synthetic data. For improving on this aspect, we investigate and compare two extensions of orthogonal popular techniques, namely plug-and-play optimization with learned priors, and a single end-to-end deep neural network trained on a larger variation of realistic synthesized training data, and compare their performance with special emphasis on model violations. We observe that the end-to-end network achieves a higher robustness and flexibility than the optimiza- tion based technique.

Lightfield imaging systems were brought to market for consumer and professional media recording. But so far, this technology is less known for applications in the industrial space. The unique optical concept developed by KLens allows to capture multiple perspectives of a scene with a single exposure as regular colour images on the camera sensor.

In this article, we propose an efficient joint image quality assessment and enhancement algorithm for the 3-D-synthesized images using a global predictor, namely, kernel ridge regression (KRR).

In this paper, we address problems associated with free-energy-principle-based image quality assessment (IQA) algorithms for objectively assessing the quality of Screen Content (SC) and three-dimensional (3-D) synthesized images and also propose a very fast and efficient IQA algorithm to address these issues.

An encoding approach based on Convolutional Neural Networks is explored to partly substitute classical heuristics-based encoder speed-ups by a systematic and automatic process. The solution allows controlling the trade-off between complexity and coding gains, in intra slices, with one single parameter.

We propose the first pixel-based JND algorithm that includes a very important component of the human vision, namely CS by measuring RMS contrast. This RMS contrast is combined with LA and CM to form a comprehensive pixel-domain model to efficiently estimate JND in the low frequency regions. For high frequency regions (edge and texture), a feedback mechanism is proposed to efficiently alleviate the over- and under-estimation of CM.

Color filter array (CFA) interpolation, or three-band demosaicking, is a process of interpolating the missing color samples in each band to reconstruct a full color image. In this paper, we are concerned with the challenging problem of multispectral demosaicking, where each band is significantly undersampled due to the increment in the number of bands.

In this paper, we propose an efficient perceptually motivated and maximum a posterior (MAP)-based generic framework for image reconstruction. This can be applied to several image/video processing applications, where there is a necessity to improve reconstruction accuracy and suppress visible artifacts, such as denoising, deinterlacing, interpolation, de-blocking of Jpeg/Jpeg-2000, and demosaicing.