CRAFT Objects from Images

By Bin Yang, Junjie Yan, Zhen Lei, S. Li. 2016

125 citations recorded by the source corpus.

Object detection is a fundamental problem in image understanding. One popular solution is the R-CNN framework [15] and its fast versions [14, 27]. They decompose the object detection problem into two cascaded easier tasks: 1) generating object proposals from images, 2) classifying proposals into various object categories. Despite that we are handling with two relatively easier tasks, they are not solved perfectly and there's still room for improvement. In this paper, we push the "divide and conquer" solution even further by dividing each task into two sub-tasks. We call the proposed method "CRAFT" (Cascade Regionproposal-network And FasT-rcnn), which tackles each task with a carefully designed network cascade. We show that the cascade structure helps in both tasks: in proposal generation, it provides more compact and better localized object proposals, in object classification, it reduces false positives (mainly between ambiguous categories) by capturing both inter-and intra-category variances. CRAFT achieves consistent and considerable improvement over the state-of the-art on object detection benchmarks like PASCAL VOC 07/12 and ILSVRC.

Subfields: Computer Vision and Pattern Recognition

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

Citations in the corpus

Cascade R-CNN: High Quality Object Detection and Instance Segmentation (2019, methodology): “[17], [18], [59] used a multi-stage procedure to generate accurate proposals, which are forwarded”
Speed/Accuracy Trade-Offs for Modern Convolutional Object Detectors (2016, background): “Since appearing in 2015, Faster R-CNN has been particularly influential, and has led to a number of follow-up works [2, 34, 33, 45, 13, 5, 19, 44, 23, 46] (including SSD and R-FCN).”
Learning Open-World Object Proposals without Learning to Classify (2021, background): “A number of follow-up works [5], [6], [23], [25] have been proposed to improve the quality of such region proposals and reduce their number in order to speed up the final detection task.”
Cascade R-CNN: Delving Into High Quality Object Detection (2017, methodology): “[36, 12, 11] used a multi-stage procedure to generate accurate proposals, and forwarded them to an accurate model (e.”
Hybrid Task Cascade for Instance Segmentation (2019, methodology): “CRAFT [44] incorporates a cascade structure into RPN [39] and Fast RCNN [16] to improve the quality of the proposal and detection results.”
Acquisition of Localization Confidence for Accurate Object Detection (2018, background): “In a series of object detectors [31, 7, 6, 21], refined boxes will be fed to the bounding box regressor again and go through the refinement for another time.”
Region Proposal by Guided Anchoring (2019, background): “There have been attempts [8, 9, 23, 31, 33, 34, 1, 2] that apply cascade architecture to reject easy samples at early layers or stages, and regress bounding boxes iteratively for progressive refinement.”
Object Detection With Deep Learning: A Review (2018, background): “detectors are built in different stages or layers [180], [220].”
Adaptive Feeding: Achieving Fast and Accurate Detections by Adaptively Combining Object Detectors (2017, unknown): “For example, a triplet combination can adds an extra model, which is more accurate but slower than those models in our experiments [11, 31], to further improve the detection accuracy without losing AF’s speed beneﬁts.”
Chained Cascade Network for Object Detection (2017, methodology): “The existing approaches in [31, 37] are used for generating RoIs in our implementation.”
Recent Advances in Deep Learning for Object Detection (2019, methodology): “For deep learning based detection algorithms, Yang et al. [153] proposed CRAFT (Cas-cade Region-proposal-network And FasT-rcnn) which learned RPN and region classiﬁers with a cascaded learning strategy.”
Learning Chained Deep Features and Classifiers for Cascade in Object Detection (2017, methodology): “The existing CNN based approaches in [33, 11] are used for generating RoIs in our implementation.”
Deep Feature Flow for Video Recognition (2016, methodology): “8% is reported in [22], by combining CRAFT [46] and DeepID-net [30] object detectors trained on the ImageNet data, using both VGG-16 [39] and GoogleNet-v2 [20] models, with various tricks (multi-scale training/testing, adding context information, model ensemble).”
Video Object Detection with an Aligned Spatial-Temporal Memory (2017, unknown): “4 T-CNN [25] DeepID+Craft [34,48] RCNN 67.”
Deep Learning for Generic Object Detection: A Survey (2018, methodology): “dding a branch to compute the scalar zoom indicator in parallel with the existing branch. There is further work attempting to generate object proposals by exploiting multilayer convolutional features [103,61,224,122]. 18 Li Liu et al. Table 5 Summarization of object proposal methods using DCNN. The ”
Reducing Network Agnostophobia (2018, background): “) causing many false positives between ambiguous object categories” [41].”
Cascade RPN: Delving into High-Quality Region Proposal Network with Adaptive Convolution (2019, background): “There have been a number of studies attempting to improve the performance of RPN [14, 37, 38, 41].”
Object Detection in 20 Years: A Survey (2019, background): “, to improve the detection of hard examples [182–184], to”
T-CNN: Tubelets With Convolutional Neural Networks for Object Detection From Videos (2016, methodology): “1) Still-Image Object Detection: Our still-image object detectors adopt the DeepID-Net [8] and CRAFT [44] frameworks and are trained with both ImageNet detection (DET) and video (VID) training datasets in ILSVRC 2015.”