Fast ANN for High‐Quality Collaborative Filtering

Tsai, Yun-Ta; Steinberger, Markus; Pajźk, Dawid; Pulli, Kari

doi:10.1111/cgf.12715

Cited by 11 publications

(3 citation statements)

References 43 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Global (nonadaptive) values for the patch size p 0 and the filtering parameter p 1 σ have been empirically investigated and reported in [1] as a function of σ . The choice of these parameters is critical when the Approximate NLM (ANLM) is used, i.e.when N is small to reduce the computational cost (e.g., N = 16 in [19]) and preserve edges better [5]. We used the procedure in Sect.…”

Section: Image Denoisingmentioning

confidence: 99%

Learning Adaptive Parameter Tuning for Image Processing

Dong¹,

Frosio²,

Kautz³

2018

View full text Add to dashboard Cite

The non-stationary nature of image characteristics calls for adaptive processing, based on the local image content. We propose a simple and flexible method to learn local tuning of parameters in adaptive image processing: we extract simple local features from an image and learn the relation between these features and the optimal filtering parameters. Learning is performed by optimizing a user defined cost function (any image quality metric) on a training set. We apply our method to three classical problems (denoising, demosaicing and deblurring) and we show the effectiveness of the learned parameter modulation strategies. We also show that these strategies are consistent with theoretical results from the literature.

show abstract

Section: Image Denoisingmentioning

confidence: 99%

Learning Adaptive Parameter Tuning for Image Processing

Dong¹,

Frosio²,

Kautz³

2018

View full text Add to dashboard Cite

show abstract

“…The self-similarity based denoising priors (BM3D, NLM, etc.) are accelerated with an approximate-nearest-neighbor (ANN) method [Tsai et al 2014]. Furthermore, we split large images into tiles and process these tiles separately to save memory.…”

Section: Performancementioning

confidence: 99%

FlexISP

Heide

Steinberger

Tsai³

et al. 2014

ACM Trans. Graph.

Self Cite

253

View full text Add to dashboard Cite

Conventional pipelines for capturing, displaying, and storing images are usually defined as a series of cascaded modules, each responsible for addressing a particular problem. While this divide-and-conquer approach offers many benefits, it also introduces a cumulative error, as each step in the pipeline only considers the output of the previous step, not the original sensor data. We propose an end-to-end system that is aware of the camera and image model, enforces natural-image priors, while jointly accounting for common image processing steps like demosaicking, denoising, deconvolution, and so forth, all directly in a given output representation (e.g., YUV, DCT). Our system is flexible and we demonstrate it on regular Bayer images as well as images from custom sensors. In all cases, we achieve large improvements in image quality and signal reconstruction compared to state-of-the-art techniques. Finally, we show that our approach is capable of very efficiently handling high-resolution images, making even mobile implementations feasible.

show abstract

“…While leveraging GPU parallelism seems obvious, in practice accelerating ANN search techniques using GPU parallelism is notoriously difficult, largely due to the memory restrictions of GPUs when compared to the amount of RAM available to CPUs. As a result, existing GPUbased methods often implement brute force approaches, are limited to small datasets of up to 225 candidate neighbors [18] or can handle only 3-dimensional vectors [16], making these approaches unsuited for many vision problems. Our method is designed to be highly parallel, and can be easily implemented on a GPU for significant improvements in query time, while even a CPU version is competitive to previous methods.…”

Section: Introductionmentioning

confidence: 99%

Efficient Large-Scale Approximate Nearest Neighbor Search on the GPU

Wieschollek

Wang

Sorkine-Hornung

et al. 2016

2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)

View full text Add to dashboard Cite

We present a new approach for efficient approximate nearest neighbor (ANN) search in high dimensional spaces, extending the idea of Product Quantization. We propose a two level product and vector quantization tree that reduces the number of vector comparisons required during tree traversal. Our approach also includes a novel highly parallelizable re-ranking method for candidate vectors by efficiently reusing already computed intermediate values.Due to its small memory footprint during traversal the method lends itself to an efficient, parallel GPU implementation. This Product Quantization Tree (PQT) approach significantly outperforms recent state of the art methods for high dimensional nearest neighbor queries on standard reference datasets. Ours is the first work that demonstrates GPU performance superior to CPU performance on high dimensional, large scale ANN problems in time-critical real-world applications, like loop-closing in videos.

show abstract

Fast ANN for High‐Quality Collaborative Filtering

Cited by 11 publications

References 43 publications

Learning Adaptive Parameter Tuning for Image Processing

Learning Adaptive Parameter Tuning for Image Processing

FlexISP

Efficient Large-Scale Approximate Nearest Neighbor Search on the GPU

Contact Info

Product

Resources

About