Exploring the representation capabilities of the HOG descriptor

Tatu, Aditya; Lauze, François; Nielsen, Mads; Kimia, Benjamin B.

doi:10.1109/iccvw.2011.6130416

Cited by 10 publications

(11 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Some of these methods look for adversarial perturbations of a source image. For instance, Tatu et al [41] show that it is possible to make any two images look nearly identical in SIFT space up to the injection of adversarial noise in the data. The complementary effect was demonstrated for CNNs by Szegedy et al [40], where an imperceptible amount of adversarial noise was shown to change the predicted class of an image to any desired class.…”

Section: Fooling Representationsmentioning

confidence: 99%

“…These show that HOG, SIFT, and early layers of CNNs are largely invertible. This apparent inconsistency may be resolved by noting that [32,40,41] require the injection of adversarial noise which is very unlikely to occur in natural images. It is not unlikely that enforcing representation to be sufficiently regular would avoid the issue.…”

Section: Fooling Representationsmentioning

confidence: 99%

See 1 more Smart Citation

Visualizing Deep Convolutional Neural Networks Using Natural Pre-images

Mahendran¹,

Vedaldi²

2016

Int J Comput Vis

462

314

View full text Add to dashboard Cite

Image representations, from SIFT and bag of visual words to convolutional neural networks (CNNs) are a crucial component of almost all computer vision systems. However, our understanding of them remains limited. In this paper we study several landmark representations, both shallow and deep, by a number of complementary visualization techniques. These visualizations are based on the concept of "natural pre-image", namely a natural-looking image whose representation has some notable property. We study in particular three such visualizations: inversion, in which the aim is to reconstruct an image from its representation, activation maximization, in which we search for patterns that maximally stimulate a representation component, and caricaturization, in which the visual patterns that a representation detects in an image are exaggerated. We pose these as a regularized energy-minimization framework and demonstrate its generality and effectiveness. In particular, we show that this method can invert representations such as HOG more accurately than recent alternatives while being applicable to CNNs too. Among our findings, we show that several layers in CNNs retain photographically accurate information about the image, with different degrees of geometric and photometric invariance.

show abstract

Section: Fooling Representationsmentioning

confidence: 99%

Section: Fooling Representationsmentioning

confidence: 99%

Visualizing Deep Convolutional Neural Networks Using Natural Pre-images

Mahendran¹,

Vedaldi²

2016

Int J Comput Vis

462

314

View full text Add to dashboard Cite

show abstract

“…Divvala et al [5] analyze part-based detectors to determine which components of object detection systems have the most impact on performance. Tatu et al [20] explored the set of images that generate identical HOG descriptors. Liu and Wang [12] designed algorithms to highlight which image regions contribute the most to a classifier's confidence.…”

Section: Related Workmentioning

confidence: 99%

HOGgles: Visualizing Object Detection Features

Vondrick

Khosla

Malisiewicz

et al. 2013

2013 IEEE International Conference on Computer Vision

239

186

View full text Add to dashboard Cite

We introduce algorithms to visualize feature spaces used by object detectors. The tools in this paper allow a human to put on 'HOG goggles' and perceive the visual world as a HOG based object detector sees it. We found that these visualizations allow us to analyze object detection systems in new ways and gain new insight into the detector's failures. For example, when we visualize the features for high scoring false alarms, we discovered that, although they are clearly wrong in image space, they do look deceptively similar to true positives in feature space. This result suggests that many of these false alarms are caused by our choice of feature space, and indicates that creating a better learning algorithm or building bigger datasets is unlikely to correct these errors. By visualizing feature spaces, we can gain a more intuitive understanding of our detection systems.

show abstract

“…A symmetric effect was demonstrated for CNNs by Szegedy et al [26], where an imperceptible amount of adversarial noise suffices to change the predicted class of an image. The apparent inconsistency is easily resolved, however, as the methods of [26,27] require the injection of high-pass structured noise which is very unlikely to occur in natural images.…”

Section: Introductionmentioning

confidence: 99%

Understanding deep image representations by inverting them

Mahendran

Vedaldi

2015

2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)

1,574

1,209

View full text Add to dashboard Cite

Image representations, from SIFT and Bag of Visual Words to Convolutional Neural Networks (CNNs), are a crucial component of almost any image understanding system. Nevertheless, our understanding of them remains limited. In this paper we conduct a direct analysis of the visual information contained in representations by asking the following question: given an encoding of an image, to which extent is it possible to reconstruct the image itself? To answer this question we contribute a general framework to invert representations. We show that this method can invert representations such as HOG and SIFT more accurately than recent alternatives while being applicable to CNNs too. We then use this technique to study the inverse of recent state-of-the-art CNN image representations for the first time. Among our findings, we show that several layers in CNNs retain photographically accurate information about the image, with different degrees of geometric and photometric invariance.

show abstract

Exploring the representation capabilities of the HOG descriptor

Cited by 10 publications

References 14 publications

Visualizing Deep Convolutional Neural Networks Using Natural Pre-images

Visualizing Deep Convolutional Neural Networks Using Natural Pre-images

HOGgles: Visualizing Object Detection Features

Understanding deep image representations by inverting them

Contact Info

Product

Resources

About