Query Adaptive Instance Search using Object Sketches

Bhattacharjee, Sreyasee Das; Yuan, Junsong; Hong, Weixiang; Ruan, Xiang

doi:10.1145/2964284.2964317

Cited by 26 publications

(21 citation statements)

References 35 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…In a recent short paper, Qi et al [26] also propose a two-branch Siamese network with contrastive loss. Their results, although comparable with other methods using shallow features, are still far behind state-of-the-art [24,6] by a large margin. As we show later, learning a single function to map disparate domains to the search space appears to under-perform designs where branch weights are learned independently or semi-independently.…”

Section: Sketchanet Forms a Component Of The Very Recent Work Ofmentioning

confidence: 52%

“…Bhattacharjee et al [24], coupled with a complex pipeline including object proposals, and query expansion. Although we also explored SketchANet, and compare with several other contemporary architectures which we show yield superior performance in a triplet framework (Sec.…”

Section: Sketchanet Forms a Component Of The Very Recent Work Ofmentioning

confidence: 99%

“…• CNN-related approaches: use deep features with various architecture settings and loss functions. These include Siamese network [26], Triplet sketch-edgemap network [31], Asymmetric feature map (AFM) [47], Quadruplet MT [35], Query-adaptive CNN with re-ranking [24].…”

Section: Benchmark Evaluationmentioning

confidence: 99%

See 2 more Smart Citations

Sketching out the details: Sketch-based image retrieval using convolutional neural networks with multi-stage regression

Bui

Ribeiro

Ponti

et al. 2018

Computers & Graphics

View full text Add to dashboard Cite

We propose and evaluate several deep network architectures for measuring the similarity between sketches and photographs, within the context of the sketch based image retrieval (SBIR) task. We study the ability of our networks to generalize across diverse object categories from limited training data, and explore in detail strategies for weight sharing, pre-processing, data augmentation and dimensionality reduction. In addition to a detailed comparative study of network configurations, we contribute by describing a hybrid multi-stage training network that exploits both contrastive and triplet networks to exceed state of the art performance on several SBIR benchmarks by a significant margin. Datasets and models are available at www.cvssp.org.

show abstract

Section: Sketchanet Forms a Component Of The Very Recent Work Ofmentioning

confidence: 52%

Section: Sketchanet Forms a Component Of The Very Recent Work Ofmentioning

confidence: 99%

See 1 more Smart Citation

Sketching out the details: Sketch-based image retrieval using convolutional neural networks with multi-stage regression

Bui

Ribeiro

Ponti

et al. 2018

Computers & Graphics

View full text Add to dashboard Cite

show abstract

“…Some previous SBIR works [153], [154], [155] use edge maps (image contours) as sketch approximation to conduct the cross-modal matching. Canny edge detector [156], Edge Boxes toolbox [157], and holistically-nested edge detection (HED) [158] are usually used to extract the edges from natural photos.…”

Section: Category and Instance Level Sbirmentioning

confidence: 99%

Deep Learning for Free-Hand Sketch: A Survey

Xu¹

2020

Preprint

View full text Add to dashboard Cite

Free-hand sketches are highly hieroglyphic and illustrative, which have been widely used by humans to depict objects or stories from ancient times to the present. The recent prevalence of touchscreen devices has made sketch creation a much easier task than ever and consequently made sketch-oriented applications increasingly more popular. The prosperity of deep learning has also immensely promoted the research for the free-hand sketch. This paper presents a comprehensive survey of the free-hand sketch oriented deep learning techniques. The main contents of this survey include: (i) The intrinsic traits and domain-unique challenges of the free-hand sketch are discussed, to clarify the essential differences between free-hand sketch and other data modalities, e.g., natural photo. (ii) The development of the free-hand sketch community in the deep learning era is reviewed, by surveying the existing datasets, research topics, and the state-of-the-art methods via a detailed taxonomy. (iii) Moreover, the bottlenecks, open problems, and potential research directions of this community have also been discussed to promote the future works.

show abstract

“…Despite the problem of object instance search has been widely exploited on image datasets [77,13,78,79,67,80] , while few works have focused on video datasets. A pioneering work for object instance search in videos is Video Google [42], which treats each video keyframe independently and ranks the videos by their best-matched keyframe.…”

Section: Fast Video Searchmentioning

confidence: 99%

Efficient visual search in images and videos

Yu¹

View full text Add to dashboard Cite

First and foremost, I would like to express my gratitude to Professor Junsong Yuan, who has supported me throughout my PhD study with his patience and knowledge, and encouraged me to explore my own way of research. During these years of studying under his supervision, I have learned a lot. I have been greatly impressed by his professional attitude towards research and teaching. I would also like to thanks Professor Kai-Kuang Ma and Professor Xueyan Tang for their valuable suggestions on various problems in my research.

show abstract

Query Adaptive Instance Search using Object Sketches

Cited by 26 publications

References 35 publications

Sketching out the details: Sketch-based image retrieval using convolutional neural networks with multi-stage regression

Sketching out the details: Sketch-based image retrieval using convolutional neural networks with multi-stage regression

Deep Learning for Free-Hand Sketch: A Survey

Efficient visual search in images and videos

Contact Info

Product

Resources

About