Pre-Attentive and Attentive Detection of Humans in Wide-Field Scenes

Object detection through search with a foveated visual system

Humans and many other species sense visual information with varying spatial resolution across the visual field (foveated vision) and deploy eye movements to actively sample regions of interests in scenes. The advantage of such varying resolution architecture is a reduced computational, hence metabolic cost. But what are the performance costs of such processing strategy relative to a scheme that processes the visual field at high spatial resolution? Here we first focus on visual search and combine object detectors from computer vision with a recent model of peripheral pooling regions found at the V1 layer of the human visual system. We develop a foveated object detector that processes the entire scene with varying resolution, uses retino-specific object detection classifiers to guide eye movements, aligns its fovea with regions of interest in the input image and integrates observations across multiple fixations. We compared the foveated object detector against a non-foveated version of the same object detector which processes the entire image at homogeneous high spatial resolution. We evaluated the accuracy of the foveated and non-foveated object detectors identifying 20 different objects classes in scenes from a standard computer vision data set (the PASCAL VOC 2007 dataset). We show that the foveated object detector can approximate the performance of the object detector with homogeneous high spatial resolution processing while bringing significant computational cost savings. Additionally, we assessed the impact of foveation on the computation of bottom-up saliency. An implementation of a simple foveated bottom-up saliency model with eye movements showed agreement in the selection of top salient regions of scenes with those selected by a non-foveated high resolution saliency model. Together, our results might help explain the evolution of foveated visual systems with eye movements as a solution that preserves perceptual performance in visual search while resulting in computational and metabolic savings to the brain.

Statistical mechanics approach to the sample deconvolution problem

In a multicellular organism different cell types express a gene in different amounts. Samples from which gene expression levels can be measured typically contain a mixture of different cell types; the resulting measurements thus give only averages over the different cell types present. Based on fluctuations in the mixture proportions from sample to sample it is in principle possible to reconstruct the underlying expression levels of each cell type: to deconvolute the sample. We use a statistical mechanics approach to the problem of deconvoluting such partial concentrations from mixed samples, explore this approach using Markov chain Monte Carlo simulations, and give analytical results for when and how well samples can be unmixed.

Adaptive Gaze Control for Object Detection

We propose a novel gaze-control model for detecting objects in images. The model, named act-detect, uses the information from local image samples in order to shift its gaze towards object locations. The model constitutes two main contributions. The first contribution is that the model’s setup makes it computationally highly efficient in comparison with existing window-sliding methods for object detection, while retaining an acceptable detection performance. act-detect is evaluated on a face-detection task using a publicly available image set. In terms of detection performance, act-detect slightly outperforms the window-sliding methods that have been applied to the face-detection task. In terms of computational efficiency, act-detect clearly outperforms the window-sliding methods: it requires in the order of hundreds fewer samples for detection. The second contribution of the model lies in its more extensive use of local samples than previous models: instead of merely using them for verifying object presence at the gaze location, the model uses them to determine a direction and distance to the object of interest. The simultaneous adaptation of both the model’s visual features and its gaze-control strategy leads to the discovery of features and strategies for exploiting the local context of objects. For example, the model uses the spatial relations between the bodies of the persons in the images and their faces. The resulting gaze control is a temporal process, in which the object’s context is exploited at different scales and at different image locations relative to the object.

Tied factor analysis for face recognition across large pose differences

Face recognition algorithms perform very unreliably when the pose of the probe face is different from the gallery face: typical feature vectors vary more with pose than with identity. We propose a generative model that creates a one-to-many mapping from an idealized "identity" space to the observed data space. In identity space, the representation for each individual does not vary with pose. We model the measured feature vector as being generated by a pose-contingent linear transformation of the identity variable in the presence of Gaussian noise. We term this model "tied" factor analysis. The choice of linear transformation (factors) depends on the pose, but the loadings are constant (tied) for a given individual. We use the EM algorithm to estimate the linear transformations and the noise parameters from training data. We propose a probabilistic distance metric which allows a full posterior over possible matches to be established. We introduce a novel feature extraction process and investigate recognition performance using the FERET, XM2VTS and PIE databases. Recognition performance compares favourably to contemporary approaches.