Multifeature-based surround inhibition improves contour detection in natural images

Learning parallel and hierarchical mechanisms for edge detection

Edge detection is one of the fundamental components of advanced computer vision tasks, and it is essential to preserve computational resources while ensuring a certain level of performance. In this paper, we propose a lightweight edge detection network called the Parallel and Hierarchical Network (PHNet), which draws inspiration from the parallel processing and hierarchical processing mechanisms of visual information in the visual cortex neurons and is implemented via a convolutional neural network (CNN). Specifically, we designed an encoding network with parallel and hierarchical processing based on the visual information transmission pathway of the "retina-LGN-V1" and meticulously modeled the receptive fields of the cells involved in the pathway. Empirical evaluation demonstrates that, despite a minimal parameter count of only 0.2 M, the proposed model achieves a remarkable ODS score of 0.781 on the BSDS500 dataset and ODS score of 0.863 on the MBDD dataset. These results underscore the efficacy of the proposed network in attaining superior edge detection performance at a low computational cost. Moreover, we believe that this study, which combines computational vision and biological vision, can provide new insights into edge detection model research.

DPED: Bio-inspired dual-pathway network for edge detection

Edge detection is significant as the basis of high-level visual tasks. Most encoder-decoder edge detection methods used convolutional neural networks, such as VGG16 or Resnet, as the encoding network. Studies on designing decoding networks have achieved good results. Swin Transformer (Swin) has recently attracted much attention in various visual tasks as a possible alternative to convolutional neural networks. Physiological studies have shown that there are two visual pathways that converge in the visual cortex in the biological vision system, and that complex information transmission and communication is widespread. Inspired by the research on Swin and the biological vision pathway, we have designed a two-pathway encoding network. The first pathway network is the fine-tuned Swin; the second pathway network mainly comprises deep separable convolution. To simulate attention transmission and feature fusion between the first and second pathway networks, we have designed a second-pathway attention module and a pathways fusion module. Our proposed method outperforms the CNN-based SOTA method BDCN on BSDS500 datasets. Moreover, our proposed method and the Transformer-based SOTA method EDTER have their own performance advantages. In terms of FLOPs and FPS, our method has more benefits than EDTER.

A visual-degradation-inspired model with HSV color-encoding for contour detection

BACKGROUND

Given energy metabolism, visual information degradation plays an essential role in the retina- lateral geniculate nucleus (LGN)-primary visual cortex (V1)-secondary visual cortex (V2) pathway, and is a pivotal issue for visual information processing. Degradation helps the visual nervous system conserve brain energy and efficiently perceive the real world even though a small fraction of visual information reaches the early visual areas. The coding of contour features (edge and corner) is achieved in the retina-LGN-V1-V2 pathway. Based on the above, we proposed a contour detection model based on degradation (CDMD).

NEW METHOD

Inspired by pupillary light reflex regulation, we took into consideration the novel approach of the hue-saturation-value (HSV) module for color encoding to meet the subtle chromaticity change rather than using the traditional red-green-blue (RGB) module, following the mechanisms of dark (DA) and light (LA) adaptation processes in photoreceptors. Meanwhile, the degradation mechanism was introduced as a novel strategy focusing only on the essential information to detect contour features, mimicking contour detection by visual perception under the restriction of axons in each optic nerve biologically. Ultimately, we employed the feedback mechanism achieving the optimal HSV value for each pixel of the experimental datasets.

RESULTS

We used the publicly available Berkeley Segmentation Data Set 500 (BSDS500) to assess the effectiveness of our CDMD model, introduced the F-measure to evaluate the results. The F-measure score was 0.65, achieved by our model. Moreover, CDMD with HSV has a better sensitivity for subtle chromaticity changes than CDMD with RGB.

COMPARISON WITH EXISTING METHODS

Experimental results demonstrated that our CDMD model, which functions close to the real visual system, achieved a more competitive performance with low computational cost than some state-of-art non-deep-learning and biologically inspired models. Compared with deep-learning-based algorithms, our model contains fewer parameters and computation time, does not require additional visual features, as well as an extra training process.

CONCLUSIONS

Our proposed CDMD model is a novel approach for contour detection, which mimics the cognitive function of contour detection in early visual areas, and realizes a competitive performance in image processing. It contributes to bridging the gap between the biological visual system and computer vision.

Hierarchical Predictive Coding-Based JND Estimation for Image Compression

The human visual system (HVS) is a hierarchical system, in which visual signals are processed hierarchically. In this paper, the HVS is modeled as a three-level communication system and visual perception is divided into three stages according to the hierarchical predictive coding theory. Then, a novel just noticeable distortion (JND) estimation scheme is proposed. In visual perception, the input signals are predicted constantly and spontaneously in each hierarchy, and neural response is evoked by the central residue and inhibited by surrounding residues. These two types' residues are regarded as the positive and negative visual incentives which cause positive and negative perception effects, respectively. In neuroscience, the effect of incentive on observer is measured by the surprise of this incentive. Thus, we propose a surprise-based measurement method to measure both perception effects. Specifically, considering the biased competition of visual attention, we define the product of the residue self-information (i.e., surprise) and the competition biases as the perceptual surprise to measure the positive perception effect. As for the negative perception effect, it is measured by the average surprise (i.e., the local Shannon entropy). The JND threshold of each stage is estimated individually by considering both perception effects. The total JND threshold is finally obtained by non-linear superposition of three stage thresholds. Furthermore, the proposed JND estimation scheme is incorporated into the codec of Versatile Video Coding for image compression. Experimental results show that the proposed JND model outperforms the relevant existing ones, and over 16% of bit rate can be reduced without jeopardizing the perceptual quality.

Learning Nonclassical Receptive Field Modulation for Contour Detection

This work develops a biologically inspired neural network for contour detection in natural images by combining the nonclassical receptive field modulation mechanism with a deep learning framework. The input image is first convolved with the local feature detectors to produce the classical receptive field responses, and then a corresponding modulatory kernel is constructed for each feature map to model the nonclassical receptive field modulation behaviors. The modulatory effects can activate a larger cortical area and thus allow cortical neurons to integrate a broader range of visual information to recognize complex cases. Additionally, to characterize spatial structures at various scales, a multiresolution technique is used to represent visual field information from fine to coarse. Different scale responses are combined to estimate the contour probability. Our method achieves state-of-the-art results among all biologically inspired contour detection models. This study provides a method for improving visual modeling of contour detection and inspires new ideas for integrating more brain cognitive mechanisms into deep neural networks.

Superpixel Segmentation Based on Anisotropic Edge Strength

Superpixel segmentation can benefit from the use of an appropriate method to measure edge strength. In this paper, we present such a method based on the first derivative of anisotropic Gaussian kernels. The kernels can capture the position, direction, prominence, and scale of the edge to be detected. We incorporate the anisotropic edge strength into the distance measure between neighboring superpixels, thereby improving the performance of an existing graph-based superpixel segmentation method. Experimental results validate the superiority of our method in generating superpixels over the competing methods. It is also illustrated that the proposed superpixel segmentation method can facilitate subsequent saliency detection.

Biologically Inspired Hierarchical Contour Detection with Surround Modulation and Neural Connection

Contour is a very important feature in biological visual cognition and has been extensively investigated as a fundamental vision problem. In connection with the limitations of conventional models in detecting image contours in complex scenes, a hierarchical image contour extraction method is proposed based on the biological vision mechanism that draws on the perceptual characteristics of the early vision for features such as edges, shapes, and colours. By simulating the information processing mechanisms of the cells’ receptive fields in the early stages of the biological visual system, we put forward a computational model that combines feedforward, lateral, and feedback neural connections to decode and obtain the image contours. Our model simulations and their results show that the established hierarchical contour detection model can adequately fit the characteristics of the biological experiment, quickly and effectively detect the salient contours in complex scenes, and better suppress the unwanted textures.

A Fast Contour Detection Model Inspired by Biological Mechanisms in Primary Vision System

Compared to computer vision systems, the human visual system is more fast and accurate. It is well accepted that V1 neurons can well encode contour information. There are plenty of computational models about contour detection based on the mechanism of the V1 neurons. Multiple-cue inhibition operator is one well-known model, which is based on the mechanism of V1 neurons' non-classical receptive fields. However, this model is time-consuming and noisy. To solve these two problems, we propose an improved model which integrates some additional other mechanisms of the primary vision system. Firstly, based on the knowledge that the salient contours only occupy a small portion of the whole image, the prior filtering is introduced to decrease the running time. Secondly, based on the physiological finding that nearby neurons often have highly correlated responses and thus include redundant information, we adopt the uniform samplings to speed up the algorithm. Thirdly, sparse coding is introduced to suppress the unwanted noises. Finally, to validate the performance, we test it on Berkeley Segmentation Data Set. The results show that the improved model can decrease running time as well as keep the accuracy of the contour detection.

ASCII Art Synthesis from Natural Photographs

While ASCII art is a worldwide popular art form, automatic generating structure-based ASCII art from natural photographs remains challenging. The major challenge lies on extracting the perception-sensitive structure from the natural photographs so that a more concise ASCII art reproduction can be produced based on the structure. However, due to excessive amount of texture in natural photos, extracting perception-sensitive structure is not easy, especially when the structure may be weak and within the texture region. Besides, to fit different target text resolutions, the amount of the extracted structure should also be controllable. To tackle these challenges, we introduce a visual perception mechanism of non-classical receptive field modulation (non-CRF modulation) from physiological findings to this ASCII art application, and propose a new model of non-CRF modulation which can better separate the weak structure from the crowded texture, and also better control the scale of texture suppression. Thanks to our non-CRF model, more sensible ASCII art reproduction can be obtained. In addition, to produce more visually appealing ASCII arts, we propose a novel optimization scheme to obtain the optimal placement of proportional-font characters. We apply our method on a rich variety of images, and visually appealing ASCII art can be obtained in all cases.

Orientation Histogram-Based Center-Surround Interaction: An Integration Approach for Contour Detection

Contour is a critical feature for image description and object recognition in many computer vision tasks. However, detection of object contour remains a challenging problem because of disturbances from texture edges. This letter proposes a scheme to handle texture edges by implementing contour integration. The proposed scheme integrates structural segments into contours while inhibiting texture edges with the help of the orientation histogram-based center-surround interaction model. In the model, local edges within surroundings exert a modulatory effect on central contour cues based on the co-occurrence statistics of local edges described by the divergence of orientation histograms in the local region. We evaluate the proposed scheme on two well-known challenging boundary detection data sets (RuG and BSDS500). The experiments demonstrate that our scheme achieves a high [Formula: see text]-measure of up to 0.74. Results show that our scheme achieves integrating accurate contour while eliminating most of texture edges, a novel approach to long-range feature analysis.

Multimodal Deep Autoencoder for Human Pose Recovery

Video-based human pose recovery is usually conducted by retrieving relevant poses using image features. In the retrieving process, the mapping between 2D images and 3D poses is assumed to be linear in most of the traditional methods. However, their relationships are inherently non-linear, which limits recovery performance of these methods. In this paper, we propose a novel pose recovery method using non-linear mapping with multi-layered deep neural network. It is based on feature extraction with multimodal fusion and back-propagation deep learning. In multimodal fusion, we construct hypergraph Laplacian with low-rank representation. In this way, we obtain a unified feature description by standard eigen-decomposition of the hypergraph Laplacian matrix. In back-propagation deep learning, we learn a non-linear mapping from 2D images to 3D poses with parameter fine-tuning. The experimental results on three data sets show that the recovery error has been reduced by 20%-25%, which demonstrates the effectiveness of the proposed method.

A biologically-inspired framework for contour detection using superpixel-based candidates and hierarchical visual cues

Contour detection has been extensively investigated as a fundamental problem in computer vision. In this study, a biologically-inspired candidate weighting framework is proposed for the challenging task of detecting meaningful contours. In contrast to previous models that detect contours from pixels, a modified superpixel generation processing is proposed to generate a contour candidate set and then weigh the candidates by extracting hierarchical visual cues. We extract the low-level visual local cues to weigh the contour intrinsic property and mid-level visual cues on the basis of Gestalt principles for weighting the contour grouping constraint. Experimental results tested on the BSDS benchmark show that the proposed framework exhibits promising performances to capture meaningful contours in complex scenes.

Boundary detection using double-opponency and spatial sparseness constraint

Brightness and color are two basic visual features integrated by the human visual system (HVS) to gain a better understanding of color natural scenes. Aiming to combine these two cues to maximize the reliability of boundary detection in natural scenes, we propose a new framework based on the color-opponent mechanisms of a certain type of color-sensitive double-opponent (DO) cells in the primary visual cortex (V1) of HVS. This type of DO cells has oriented receptive field with both chromatically and spatially opponent structure. The proposed framework is a feedforward hierarchical model, which has direct counterpart to the color-opponent mechanisms involved in from the retina to V1. In addition, we employ the spatial sparseness constraint (SSC) of neural responses to further suppress the unwanted edges of texture elements. Experimental results show that the DO cells we modeled can flexibly capture both the structured chromatic and achromatic boundaries of salient objects in complex scenes when the cone inputs to DO cells are unbalanced. Meanwhile, the SSC operator further improves the performance by suppressing redundant texture edges. With competitive contour detection accuracy, the proposed model has the additional advantage of quite simple implementation with low computational cost.

Potential roles of the interaction between model V1 neurons with orientation-selective and non-selective surround inhibition in contour detection

Both the neurons with orientation-selective and with non-selective surround inhibition have been observed in the primary visual cortex (V1) of primates and cats. Though the inhibition coming from the surround region (named as non-classical receptive field, nCRF) has been considered playing critical role in visual perception, the specific role of orientation-selective and non-selective inhibition in the task of contour detection is less known. To clarify above question, we first carried out computational analysis of the contour detection performance of V1 neurons with different types of surround inhibition, on the basis of which we then proposed two integrated models to evaluate their role in this specific perceptual task by combining the two types of surround inhibition with two different ways. The two models were evaluated with synthetic images and a set of challenging natural images, and the results show that both of the integrated models outperform the typical models with orientation-selective or non-selective inhibition alone. The findings of this study suggest that V1 neurons with different types of center–surround interaction work in cooperative and adaptive ways at least when extracting organized structures from cluttered natural scenes. This work is expected to inspire efficient phenomenological models for engineering applications in field of computational machine-vision.