Object recognition with features inspired by visual cortex

By Thomas Serre, Lior Wolf, Tomaso A. Poggio. 2005

998 citations recorded by the source corpus.

Subfields: Cognitive Neuroscience, 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 Cognitive Neuroscience

In Computer Vision and Pattern Recognition

Citations in the corpus

High-Performance Neural Networks for Visual Object Classification (2011, positive): “…studies of the visual cortex justify the use of such ﬁlters in the so-called standard model for object recognition (Riesenhuber and Poggio, 1999; Serre et al. , 2007; Mutch and Lowe, 2008), whose ﬁlters are ﬁxed, in contrast to those of Convolutional Neural Networks (CNNs) (LeCun et al. ,…”
Unstructured human activity detection from RGBD images (2011, methodology): “We have also made the dataset and code available open-source at: http://pr.cs.cornell.edu/humanactivities”
A Biologically Inspired System for Action Recognition (2007, methodology): “Motivated by the recent success of biologically inspired approaches for the recognition of objects in real-world applications [25, 16, 20], we here extend a neurobiological”
A quantitative theory of immediate visual recognition. (2007, methodology): “We used a one-versus-all approach, training eight binary classifiers , one for each category against the rest of the categories, and then taking the classifier prediction to be the maximum among the eight classifiers (for further details, see Hung et al., 2005; Serre et al., 2005).”
Demographic Estimation from Face Images: Human vs. Machine Performance (2015, methodology): “Following the success of biologically inspired features [58] in object detection and classification [59], face recognition [60], and automatic age estimation [23], [25], [33], we choose to use BIF.”
Deep networks for robust visual recognition (2010, methodology): “Sparsely connected weights have been widely used in visual recognition algorithms (Fukushima, 1983; Le- Cun et al., 1998; Serre et al., 2005).”
SVM-KNN: Discriminative Nearest Neighbor Classification for Visual Category Recognition (2006, methodology): “Correctness rate of Algorithm B (plotted the same format as in [42] and [15]), best viewed in color, Results from this work and others: Lazebnik, Schmid & Ponce [18], Berg [2], Mutch & Lowe [27], Grauman & Darrell [15], Berg, Berg & Malik [3], Wang, Zhang & Fei-Fei [42], Holub, Welling & Perona [17]”
Kernel Analysis of Deep Networks (2011, background): “Similarly, in the context of sequential data, we can postulate that dedicating the early layers of the architecture to a convolutional preprocessing is also a more effective (LeCun, 1989; Serre et al., 2005) and biologically plausible (Ringach, 2002) way of solving the learning problem.”
Towards Scalable Representations of Object Categories: Learning a Hierarchy of Parts (2007, methodology): “The authors of [16] use predesigned filters and process the visual information in the feed-forward manner, while their recent version [19] exchanged the intermediate layer with random combinations of local edge arrangements rather than choosing the features in accordance with the natural statistics.”
Unsupervised Learning of Invariant Feature Hierarchies with Applications to Object Recognition (2008, methodology): “When multiple such levels are stacked, the resulting architecture is essentially identical to the Neocognitron [7], the Convolutional Network [13] [10], and the HMAX, or so-called " Standard Model " architecture [20] [17].”
Human Activity Detection from RGBD Images (2011, methodology): “Hierarchical techniques for activity recognition have been used as well, but these typically focus on neurologically-inspired visual cortex-type models (Giese and Poggio, 2003; Serre, Wolf, and Poggio, 2005; Mutch and Lowe, 2006; Ran-zato et al., 2007).”
Representational Similarity Analysis – Connecting the Branches of Systems Neuroscience (2008, positive): “Serre et al. (2005) builds on the complex-cell outputs of the V1 model described above (implemented by the same group).”
Human vs. Computer in Scene and Object Recognition (2014, methodology): “…designed for scene classification (e.g., GIST [10]: the output energy of a bank of 24 Gabor-like filters tuned to 8 orientations at 4 different scales on a 4 × 4 grid; similar core computation for gistPadding) while others are proposed for object recognition (e.g. HOG [4], HMAX [22], LBP [9]).”
Automated identification of animal species in camera trap images (2013, background): “Spatial pyramid matching (SPM) with max pooling [11] can not only model the spatial layout of local image features, but also achieve translation invariance of animal body.”
A Theoretical Analysis of Feature Pooling in Visual Recognition (2010, methodology): “, 2009), the HMAX class of models which uses max pooling (Serre et al., 2005), and some models of the primary visual cortex area V1 (Pinto et al.”
Evaluation of Pooling Operations in Convolutional Architectures for Object Recognition (2010, methodology): “In the HMAX model [21], each complex cell receives input from a distinct patch of a simple cell map.”