Combining Bottom-Up, Top-Down, and Smoothness Cues for Weakly Supervised Image Segmentation

Kim

2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

et al. 2019

454

359

The main obstacle to weakly supervised semantic image segmentation is the difficulty of obtaining pixel-level information from coarse image-level annotations. Most methods based on image-level annotations use localization maps obtained from the classifier, but these only focus on the small discriminative parts of objects and do not capture precise boundaries. FickleNet explores diverse combinations of locations on feature maps created by generic deep neural networks. It selects hidden units randomly and then uses them to obtain activation scores for image classification. Fick-leNet implicitly learns the coherence of each location in the feature maps, resulting in a localization map which identifies both discriminative and other parts of objects. The ensemble effects are obtained from a single network by selecting random hidden unit pairs, which means that a variety of localization maps are generated from a single image. Our approach does not require any additional training steps and only adds a simple layer to a standard convolutional neural network; nevertheless it outperforms recent comparable techniques on the Pascal VOC 2012 benchmark in both weakly and semi-supervised settings.

Section: Comparison To the State Of The Artmentioning

confidence: 99%

FickleNet: Weakly and Semi-Supervised Semantic Image Segmentation Using Stochastic Inference

Kim

2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

et al. 2019

454

359

2019 IEEE/CVF International Conference on Computer Vision Workshop (ICCVW)

“…Oh et al [29] and Chaudhry et al [5] considered linking saliency and attention cues together, but they adopted different strategies to acquire semantic objects. Roy and Todorovic [36] leveraged both bottom-up and top-down attention cues and fused them via a conditional random field as a recurrent network. Ahn et al [1] use image level class labels to generate an initial set of CAMs and then propagate those CAMs by using random walk predictions from AffinityNet.…”

Section: Related Workmentioning

confidence: 99%

Harvesting Information from Captions for Weakly Supervised Semantic Segmentation

Sawatzky

Banerjee

Gall

2019

Since acquiring pixel-wise annotations for training convolutional neural networks for semantic image segmentation is time-consuming, weakly supervised approaches that only require class tags have been proposed. In this work, we propose another form of supervision, namely image captions as they can be found on the Internet. These captions have two advantages. They do not require additional curation as it is the case for the clean class tags used by current weakly supervised approaches and they provide textual context for the classes present in an image. To leverage such textual context, we deploy a multi-modal network that learns a joint embedding of the visual representation of the image and the textual representation of the caption. The network estimates text activation maps (TAMs) for class names as well as compound concepts, i.e. combinations of nouns and their attributes. The TAMs of compound concepts describing classes of interest substantially imrpove the quality of the estimated class activation maps which are then used to train a network for semantic segmentation. We evaluate our method on the COCO dataset where it achieves state of the art results for weakly supervised image segmentation.

“…Our method achieves mIoU values of 63.9 [45] 49.8 51.2 TransferNet CVPR '16 [11] 52.1 51.2 AISI ECCV '18 [16] 61.3 62. [33] 52.8 53.7 TPL ICCV '17 [22] 53.1 53.8 AE_PSL CVPR '17 [44] 55.0 55.7 DCSP BMVC '17 [2] 58.6 59.2 MEFF CVPR '18 [9] -55.6 GAIN CVPR '18 [26] 55.3 56.8 MCOF CVPR '18 [43] 56.2 57.6 AffinityNet CVPR '18 [1] 58.4 60.5 DSRG CVPR '18 [17] 59.0 60.4 MDC CVPR '18 [46] 60.4 60.8 SeeNet NIPS '18 [15] 61.1 60.7 FickleNet CVPR '19 [24] 61.2 61.9 Ours 63.9 65.0 and 65.0 for PASCAL VOC 2012 validation and test images respectively, which is 94.4% of that of DeepLab [3], trained with fully annotated data, which achieved an mIoU of 67.6 on validation images. Our method is 3.1% better on test images than the best method which uses only image-level annotations for supervision.…”

Section: Results On Image Segmentationmentioning

confidence: 99%

Frame-to-Frame Aggregation of Active Regions in Web Videos for Weakly Supervised Semantic Segmentation

Kim

2019 IEEE/CVF International Conference on Computer Vision (ICCV)

et al. 2019

When a deep neural network is trained on data with only image-level labeling, the regions activated in each image tend to identify only a small region of the target object. We propose a method of using videos automatically harvested from the web to identify a larger region of the target object by using temporal information, which is not present in the static image. The temporal variations in a video allow different regions of the target object to be activated. We obtain an activated region in each frame of a video, and then aggregate the regions from successive frames into a single image, using a warping technique based on optical flow. The resulting localization maps cover more of the target object, and can then be used as proxy ground-truth to train a segmentation network. This simple approach outperforms existing methods under the same level of supervision, and even approaches relying on extra annotations. Based on VGG-16 and ResNet 101 backbones, our method achieves the mIoU of 65.0 and 67.4, respectively, on PASCAL VOC 2012 test images, which represents a new state-of-the-art.