Color Constancy Algorithms for Object and Face Recognition

Kanan, Christopher; Flores, Arturo; Cottrell, Garrison W.

doi:10.1007/978-3-642-17289-2_20

Cited by 9 publications

(5 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The dataset consists of eight classes; bark, boat, bike, graffiti, wall, trees, leuven and ubc. Amsterdam Library of Object Images (ALOI) dataset [48] also had been used by most researchers [49,50]. The dataset consists of one-thousand small objects classes.…”

Section: Methodsmentioning

confidence: 99%

Distinctive accuracy measurement of binary descriptors in mobile augmented reality

2019

View full text Add to dashboard Cite

Mobile Augmented Reality (MAR) requires a descriptor that is robust to changes in viewing conditions in real time application. Many different descriptors had been proposed in the literature for example floating-point descriptors (SIFT and SURF) and binary descriptors (BRIEF, ORB, BRISK and FREAK). According to literature, floating-point descriptors are not suitable for real-time application because its operating speed does not satisfy real-time constraints. Binary descriptors have been developed with compact sizes and lower computation requirements. However, it is unclear which binary descriptors are more appropriate for MAR. Hence, a distinctive and efficient accuracy measurement of four state-of-the-art binary descriptors, namely, BRIEF, ORB, BRISK and FREAK were performed using the Mikolajczyk dataset and ALOI dataset to identify the most appropriate descriptor for MAR in terms of computation time and robustness to brightness, scale and rotation changes. The obtained results showed that FREAK is the most appropriate descriptor for MAR application as it able to produce an application that are efficient (shortest computation time) and robust towards scale, rotation and brightness changes.

show abstract

Section: Methodsmentioning

confidence: 99%

Distinctive accuracy measurement of binary descriptors in mobile augmented reality

2019

View full text Add to dashboard Cite

show abstract

“…Because SIFT descriptors are sensitive to an image's gamma encoding [19], we applied a retina-like nonlinear brightness normalization procedure to the image [20]. This is given by…”

Section: Image Featuresmentioning

confidence: 99%

Fine-grained object recognition with Gnostic Fields

Kanan

2014

IEEE Winter Conference on Applications of Computer Vision

Self Cite

View full text Add to dashboard Cite

Much object recognition research is concerned with basic-level classification, in which objects differ greatly in visual shape and appearance, e.g., desk vs duck. In contrast, fine-grained classification involves recognizing objects at a subordinate level, e.g., Wood duck vs Mallard duck. At the basic-level objects tend to differ greatly in shape and appearance, but these differences are usually much more subtle at the subordinate level, making fine-grained classification especially challenging. In this work, we show that Gnostic Fields, a brain-inspired model of object categorization, excel at fine-grained recognition. dataset of 30.5% over the state-of-theart and a 25.5% relative improvement on the Stanford Dogs dataset. We also demonstrate that Gnostic Fields can be sped up, enabling real-time classification in less than 70 ms per image. Gnostic Fields exceeded state-of-the-art methods on benchmark bird classification and dog breed recognition datasets, achieving a relative improvement on the Caltech-UCSD

show abstract

“…Subsequently, we apply a cone-like nonlinearity to the LMS pixels. This preprocessing helps the model cope with large-scale changes in brightness [6,10,14], and it is related to gamma correction [15]. The formulation we use is given by…”

Section: Cone-likementioning

confidence: 99%

Active Object Recognition with a Space-Variant Retina

Kanan

2013

ISRN Machine Vision

Self Cite

View full text Add to dashboard Cite

When independent component analysis (ICA) is applied to color natural images, the representation it learns has spatiochromatic properties similar to the responses of neurons in primary visual cortex. Existing models of ICA have only been applied to pixel patches. This does not take into account the space-variant nature of human vision. To address this, we use the space-variant logpolar transformation to acquire samples from color natural images, and then we apply ICA to the acquired samples. We analyze the spatiochromatic properties of the learned ICA filters. Qualitatively, the model matches the receptive field properties of neurons in primary visual cortex, including exhibiting the same opponent-color structure and a higher density of receptive fields in the foveal region compared to the periphery. We also adopt the "self-taught learning" paradigm from machine learning to assess the model's efficacy at active object and face classification, and the model is competitive with the best approaches in computer vision.

show abstract

Color Constancy Algorithms for Object and Face Recognition

Cited by 9 publications

References 20 publications

Distinctive accuracy measurement of binary descriptors in mobile augmented reality

Distinctive accuracy measurement of binary descriptors in mobile augmented reality

Fine-grained object recognition with Gnostic Fields

Active Object Recognition with a Space-Variant Retina

Contact Info

Product

Resources

About