Visual Recognition Performance Research Articles

Experimental study on effect of pavement background on obstacle visibility in LED lighting environment of road tunnel

The tunnel pavement is generally made of asphalt or concrete. Due to the relatively fixed material of pavement, the effect of tunnel pavement setting on the lighting environment and visual performance of drivers has not received sufficient attention, especially the impact on the visual performance of drivers during driving has not been revealed. Therefore, experimental research on the visual recognition performance of an obstacle on asphalt and concrete pavements inside tunnels during dynamic driving was conducted in this study. The results indicate that under the same pavement illumination, the luminance on concrete pavement is higher than that on asphalt pavement due to the higher reflectance of concrete. The visible distance of the human eyes for a gray obstacle with a reflectance of 0.2 on the concrete pavement is greater than that on the asphalt pavement, and the visible distance of the obstacle on the concrete pavement increases by more than 28%. When the color of the obstacle and the pavement are close, it can be challenging for observers to recognize the obstacle, and the pavement and obstacle need to have a higher level of luminance for the recognition. During dynamic driving, the visible distance at a speed of 60 km/h is 1.2 to 1.4 times that at a speed of 80 km/h, which means the influence of vehicle speed on the human eye’s recognition of obstacles on asphalt and concrete pavements should be taken into consideration in the design of road tunnel lighting. The correlated color temperature and S/P value of LED light have little impact on the visible distance of human eyes to the obstacle on the asphalt and concrete pavements, but it does create different visual perceptions. As the correlated color temperature and S/P value increase, the lighting environment of the tunnel gradually gives a brighter feeling to drivers.

Did H.M. exhibit accelerated long-term forgetting? Measuring forgetting in amnesia

The early investigations of patient H.M. inaugurated the modern era of memory research. During the 1970s and 1980s, a key debate over whether H.M. with bilateral medial temporal lobe lesions exhibited accelerated long-term forgetting attracted an increasing interest in forgetting research among amnestic patients. Huppert and Piercy (1979) examined H.M.’s performance in visual recognition at 10-minute, 1-day, and 7-day intervals and suggested that H.M. was subjected to rapid forgetting compared with Korsakoff patients and healthy participants reported in Huppert and Piercy (1978). In contrast, Freed et al. (1987) employed the same experimental paradigm and concluded that forgetting rates in H.M. did not differ from those in healthy controls. These incompatible findings highlighted a methodological challenge in measuring forgetting in the cross-group comparison design, where closely equalising the initial performance between patient and control groups is usually suggested. The re-analysis in this viewpoint, using both linear- and nonlinear-based modelling, reconciled the discrepancy between the aforementioned studies. Our results indicated that the rate of forgetting in H.M. did not differ from that in healthy controls, regardless of whether the initial performance was closely matched. Here, we suggest that the cross-group comparisons in forgetting studies do not necessarily seek a perfect match in initial performance unless the risks of confounding encoding and retrieval processes can be effectively controlled.

HeLP: The Hebrew Lexicon project

Lexicon projects (LPs) are large-scale data resources in different languages that present behavioral results from visual word recognition tasks. Analyses using LP data in multiple languages provide evidence regarding cross-linguistic differences as well as similarities in visual word recognition. Here we present the first LP in a Semitic language—the Hebrew Lexicon Project (HeLP). HeLP assembled lexical decision (LD) responses to 10,000 Hebrew words and nonwords, and naming responses to a subset of 5000 Hebrew words. We used the large-scale HeLP data to estimate the impact of general predictors (lexicality, frequency, word length, orthographic neighborhood density), and Hebrew-specific predictors (Semitic structure, presence of clitics, phonological entropy) of visual word recognition performance. Our results revealed the typical effects of lexicality and frequency obtained in many languages, but more complex impact of word length and neighborhood density. Considering Hebrew-specific characteristics, HeLP data revealed better recognition of words with a Semitic structure than words that do not conform to it, and a drop in performance for words comprising clitics. These effects varied, however, across LD and naming tasks. Lastly, a significant inhibitory effect of phonological ambiguity was found in both naming and LD. The implications of these findings for understanding reading in a Semitic language are discussed.

Skeuomorphic or flat? The effects of icon style on visual search and recognition performance

Although there have been many previous studies on icon visual search and recognition performance, only a few have considered the effects of both the internal and external characteristics of icons. In this behavioral study, we employed a visual search task and a semantic recognition task to explore the effects of icon style, semantic distance (SD), and task difficulty on users’ performance in perceiving and identifying icons. First, we created and filtered 64 new icons, which were divided into four different groups (flat design & close SD, flat design & far SD, skeuomorphic design & close SD, skeuomorphic design & far SD) through expert evaluation. A total of 40 participants (13 men and 27 women, ages ranging from 19 to 25 years, mean age = 21.9 years, SD=1.93) were asked to perform an icon visual search task and an icon recognition task after a round of learning. Participants’ accuracy and response time were measured as a function of the following independent variables: two icon styles (flat or skeuomorphic style), two levels of SD (close or far), and two levels of task difficulty (easy or difficult). The results showed that flat icons had better visual search performance than skeuomorphic icons; this beneficial effect increased as the task difficulty increased. However, in the icon recognition task, participants’ performance in recalling skeuomorphic icons was significantly better than that in recalling flat icons. Furthermore, a strong interaction effect between icon style and task difficulty was observed for response time. As the task difficulty decreased, the difference in recognition performance between these two different icon styles increased significantly. These findings provide valuable guidance for the design of icons in human–computer interaction interfaces.

Dual low-rank structure embedding for robust visual information processing

The low-rank (LR) property is widely applied to capture the global as well as intrinsic structure of the given data in different visual information processing tasks. Actually, there are three key information to determine the performance and generalization low-rank property based methods: (1) visual intrinsic structural information, (2) visual representation structural information, (3) visual robust information. To achieve these jointly in a unified framework, in this paper, we propose Dual Low-rank Structure Embedding (DLSE) that embeds structural and robust information. We additionally proposed the Joint Matrix-based Linear Representation (JMLR) and theoretically proved it can realize DLSE. The proposed method was validated on 6 datasets (from 1,440 samples to 70,000 samples in size) and showed a promising performance in visual recognition (14.32% improvement compared with the deep features). In addition, we performed multiple analysis on robustness, representation, and its parameters to show the effectiveness of DLSE from different aspects.

Analyzing to discover origins of CNNs and ViT architectures in medical images

In this paper, we introduce in-depth the analysis of CNNs and ViT architectures in medical images, with the goal of providing insights into subsequent research direction. In particular, the origins of deep neural networks should be explainable for medical images, but there has been a paucity of studies on such explainability in the aspect of deep neural network architectures. Therefore, we investigate the origin of model performance, which is the clue to explaining deep neural networks, focusing on the two most relevant architectures, such as CNNs and ViT. We give four analyses, including (1) robustness in a noisy environment, (2) consistency in translation invariance property, (3) visual recognition with obstructed images, and (4) acquired features from shape or texture so that we compare origins of CNNs and ViT that cause the differences of visual recognition performance. Furthermore, the discrepancies between medical and generic images are explored regarding such analyses. We discover that medical images, unlike generic ones, exhibit class-sensitive. Finally, we propose a straightforward ensemble method based on our analyses, demonstrating that our findings can help build follow-up studies. Our analysis code will be publicly available.

Service performance evaluation of light-transmitting concrete lane markings on highways based on accelerated pavement testing

Applying light-transmitting concrete (LTC) to highway lane markings is a novel idea to increase the visual recognition distance and promote traffic safety. However, compared to its excellent visual recognition performance, its service performance has received little attention in the current research. This paper conducted an accelerated pavement test to investigate the behaviour of rutting depth, British Pendulum Number (BPN) and water permeability coefficient of the LTC lane markings. The results show that rutting depth of LTC is lower than that of asphalt pavement. As a result, small height steps appeared at the seam of the two parts. For the anti-sliding performance represented by the BPN, the BPN of LTC gradually exceeds asphalt pavement and maintains a slightly higher value. For the water permeability coefficient, the LTC is almost un-permeable, avoiding water damage during service. According to these findings, the LTC lane markings meet the requirements of highway service performance. .

Next generation of computer vision for plant disease monitoring in precision agriculture: A contemporary survey, taxonomy, experiments, and future direction

Efficient and rational monitoring of plant health is an essential prerequisite for ensuring optimal crop production and resource management in the field of agriculture. Computer vision techniques have revolutionized visual disease monitoring with their exceptional visual recognition performance. However, despite the outstanding results, the widespread acceptance of these methods in agriculture practice is still in its early stages. This study presents a comprehensive survey of the next generation of computer vision models applied to plant disease monitoring in precision agriculture. Our study begins by tracing the evolution of agricultural computer vision research over the past decade, encompassing legacy methods such as convolutional neural networks (CNNs), progressing to newer techniques like vision transformers (ViTs), and culminating in cutting-edge vision multi-layer perceptrons (MLPs). Next, our study embraces both qualitative and quantitative approaches, supporting a profound review of literature and classifying methodologies and experimental approaches. A significant contribution lies in our comprehensive taxonomy, offering a fine-grained categorization of current computer vision models. This taxonomy meticulously highlights the potentials and limitations of these models while explaining their roles in improving plant disease management. Moreover, extensive experimental comparisons are conducted on PlantVillage dataset to evaluate the performance of state-of-the-art computer-vision models for plant recognition data. The obtained results are then utilized to draw insightful conclusions about the behavior of these models and provide guidance for selecting the most suitable one for specific tasks at hand. Additionally, we discuss open research avenues and future directions of computer-vision models in plant disease management including challenges related to the data scarcity, the computational efficiency, need for explainability, and multi-modal analysis.

Feasibility study of functional near-infrared spectroscopy in the ventral visual pathway for real-life applications.

fNIRS-based neuroenhancement depends on the feasible detection of hemodynamic responses in target brain regions. Using the lateral occipital complex (LOC) and the fusiform face area (FFA) in the ventral visual pathway as neurofeedback targets boosts performance in visual recognition. However, the feasibility of utilizing fNIRS to detect LOC and FFA activity in adults remains to be validated as the depth of these regions may exceed the detection limit of fNIRS. This study aims to investigate the feasibility of using fNIRS to measure hemodynamic responses in the ventral visual pathway, specifically in the LOC and FFA, in adults. We recorded the hemodynamic activities of the LOC and FFA regions in 35 subjects using a portable eight-channel fNIRS instrument. A standard one-back object and face recognition task was employed to elicit selective brain responses in the LOC and FFA regions. The placement of fNIRS optodes for LOC and FFA detection was guided by our group's transcranial brain atlas (TBA). Our findings revealed selective activation of the LOC target channel (CH2) in response to objects, whereas the FFA target channel (CH7) did not exhibit selective activation in response to faces. Our findings indicate that, although fNIRS detection has limitations in capturing FFA activity, the LOC region emerges as a viable target for fNIRS-based detection. Furthermore, our results advocate for the adoption of the TBA-based method for setting the LOC target channel, offering a promising solution for optrode placement. This feasibility study stands as the inaugural validation of fNIRS for detecting cortical activity in the ventral visual pathway, underscoring its ecological validity. We suggest that our findings establish a pivotal technical groundwork for prospective real-life applications of fNIRS-based research.

An Adaptable Model for Medical Image Classification Using the Streamlined Attention Mechanism

The resurgence of deep learning has improved computer vision by increasing its applicability and scalability for challenges in the real world. Specifically, utilizing attention in computer vision tasks has improved the performance of models to a superior level. Today we need medical diagnostic tools to tackle the immediate needs of the population suffering from Cancer and COVID-19. Thus, an end-to-end screening is tedious and needs validation expertise. But, with the current deep learning methods, it is quite possible to provide a diagnostic tool that can assist doctors and patients with their immediate needs. So, to tackle these real-world problems, we have proposed a method that is implied on 3 diverse standard datasets in the field of medical imagery, which are Skin Lesions, Brain tumors, and COVID-19 classification. To justify the model’s performance, the authors have experimented on 5 diverse data sets ranging from binary class to multi-class. The experimentation has shown that the proposed “streamlined-attention module” is not only capable of producing superior performance in fine-grained visual recognition but also bio-medical imagery. To further justify, we have illustrated Grad-Cam heat map visualizations to the model and show that it can extract the detailed features with proportionate attention. Our results have proved that our methods excel in their performance compared to that of existing methods. It has achieved state-of-the-art accuracy scores on COVID-19 and HAM10000 datasets with a well-guided explainable result. This work represents a significant advancement in the field of medical image processing with clear results, and the authors anticipate that the suggested method will prove to be a useful tool for medical professionals in the detection and identification of diseases like Covid-19 and cancer. This paper provided the best accuracy for COVID-19 multiclass (94.75 ± 1.07) and HAM10000 (94.31 ± 0.91).

DBN-Mix: Training dual branch network using bilateral mixup augmentation for long-tailed visual recognition

There is growing interest in the challenging visual perception task of learning from long-tailed class distributions. The extreme class imbalance in the training dataset biases the model to prefer recognizing majority class data over minority class data. Furthermore, the lack of diversity in minority class samples makes it difficult to find a good representation. In this paper, we propose an effective data augmentation method, referred to as bilateral mixup augmentation, which can improve the performance of long-tailed visual recognition. The bilateral mixup augmentation combines two samples generated by a uniform sampler and a re-balanced sampler and augments the training dataset to enhance the representation learning for minority classes. We also reduce the classifier bias using class-wise temperature scaling, which scales the logits differently per class in the training phase. We apply both ideas to the dual-branch network (DBN) framework, presenting a new model, named dual-branch network with bilateral mixup (DBN-Mix). Experiments on popular long-tailed visual recognition datasets show that DBN-Mix improves performance significantly over baseline and that the proposed method achieves state-of-the-art performance in some categories of benchmarks.

4 Assessing Visuospatial Skills in Parkinson’s Disease Using the Identi-Fi

Objective:Patients with Parkinson’s disease (PD) commonly show deficits on tests of visuospatial functioning. The Identi-Fi is a new measure of visual organization and recognition composed of two components. The Visual Recognition (VR) subtest asks persons to identify an object that has been broken its pieces and rearranged, akin to the Hooper Visual Organization Test, but using updated and colorful pictures. The Visual Matching (VM) subtest involves showing the same stimuli, but the examinee must select the correct response from among five choices (1 correct and 4 foils), placing greater demand on visuospatial discrimination. Together, the two subtests comprise the Visual Organization Index (VOI), reflecting overall visual processing and organization ability. The present study examined performance on the Identi-Fi in patients with PD and its association with other aspects of cognition.Participants and Methods:Participants were 23 patients with PD (95% male; mean age = 69.7 years [SD = 7.8], range = 47-79) and 12 patients with cognitive concerns (CC) who were intact on neuropsychological testing (excluding consideration of Identi-Fi scores; 50% male, mean age = 71.08 [SD = 6.27], range = 60-78) seen for a neuropsychological evaluation at a large Northeastern medical center. As part of a larger battery, patients completed the Identi-Fi, Trail Making Test (TMT), Category Fluency, Test of Premorbid Functioning (TOPF), and Brief Visuospatial Memory Test, Revised (BVMT-R).Results:The PD group performed significantly worse than the CC group on VR and VM, as well as VOI, of the Identi-Fi (p < .001). Within the PD group, poorer VR, VM, and VOI performance was associated with lower scores on the TOPF (p < .05), BVMT-R learning (p < .05) and delayed recall (p < .05), as well as TMT Parts A and B (p < .05). VR was significantly correlated with Category Fluency (p < .05), while a trend was seen for the association between VOI and Category Fluency (p = .094).Conclusions:Identi-Fi performance was worse in the PD group than the CC group, which is consistent with prior research indicating that visuospatial processing is often abnormal in patients with PD. Furthermore, findings indicate that poorer performance on the Identi-Fi in the PD group is associated with poorer cognitive functioning in other domains (i.e., visuospatial learning and memory, processing speed, cognitive flexibility, and semantic fluency), as well as lower premorbid intellectual functioning. While these findings suggest that the Identi-Fi is useful in identifying visuospatial dysfunction in PD, findings should be interpreted with caution given the small sample sizes and uneven gender distribution

An improved YOLO v4 used for grape detection in unstructured environment.

Visual recognition is the most critical function of a harvesting robot, and the accuracy of the harvesting action is based on the performance of visual recognition. However, unstructured environment, such as severe occlusion, fruits overlap, illumination changes, complex backgrounds, and even heavy fog weather, pose series of serious challenges to the detection accuracy of the recognition algorithm. Hence, this paper proposes an improved YOLO v4 model, called YOLO v4+, to cope with the challenges brought by unstructured environment. The output of each Resblock_body in the backbone is processed using a simple, parameterless attention mechanism for full dimensional refinement of extracted features. Further, in order to alleviate the problem of feature information loss, a multi scale feature fusion module with fusion weight and jump connection structure was pro-posed. In addition, the focal loss function is adopted and the hyperparameters α, γ are adjusted to 0.75 and 2. The experimental results show that the average precision of the YOLO v4+ model is 94.25% and the F1 score is 93%, which is 3.35% and 3% higher than the original YOLO v4 respectively. Compared with several state-of-the-art detection models, YOLO v4+ not only has the highest comprehensive ability, but also has better generalization ability. Selecting the corresponding augmentation method for specific working condition can greatly improve the model detection accuracy. Applying the proposed method to harvesting robots may enhance the applicability and robustness of the robotic system.

Multi-proxy feature learning for robust fine-grained visual recognition

Visual representation for fine-grained visual recognition can be learned by mandatorily enforcing all samples of the same category into a uniform representation. This strict training objective performs well under closed-set setting but is not applicable to data in the wild containing noisy annotations and long-tailed distributions, e.g., it may lead to a feature space biased to head categories. This paper tackles this challenge by pursuing a more balanced and discriminative feature space by first retaining intra-class variances to isolate noises, then eliminating intra-class variances to improve the visual recognition performance. We propose the Compact Memory Updater to maintain a memory bank, which memorizes proxy features to represent multiple typical appearances of each category in the training set. The Proxy-based Feature Enhancement hence leverages proxy features to ensure samples of the same category have similar features. Iteratively running those two modules boosts the robustness and discriminative power of the learnt representation, hence facilitates various fine-grained visual recognition tasks including person re-identification (re-id), image classification and retrieval. Extensive experiments on noisy and long-tailed training sets show this Multi-Proxy Feature Learning (MPFL) framework achieves promising performance. For instance on a training set with 90% one-shot categories, MPFL outperforms the recent long-tailed person re-id method LEAP-AF by 16.9% in rank-1 accuracy.

Learning Semantic Degradation-Aware Guidance for Recognition-Driven Unsupervised Low-Light Image Enhancement

Low-light images suffer severe degradation of low lightness and noise corruption, causing unsatisfactory visual quality and visual recognition performance. To solve this problem while meeting the unavailability of paired datasets in wide-range scenarios, unsupervised low-light image enhancement (ULLIE) techniques have been developed. However, these methods are primarily guided to alleviate the degradation effect on visual quality rather than semantic levels, hence limiting their performance in visual recognition tasks. To this end, we propose to learn a Semantic Degradation-Aware Guidance (SDAG) that perceives the low-light degradation effect on semantic levels in a self-supervised manner, which is further utilized to guide the ULLIE methods. The proposed SDAG utilizes the low-light degradation factors as augmented signals to degrade the low-light images, and then capture their degradation effect on semantic levels. Specifically, our SDAG employs the subsequent pre-trained recognition model extractor to extract semantic representations, and then learns to self-reconstruct the enhanced low-light image and its augmented degraded images. By constraining the relative reconstruction effect between the original enhanced image and the augmented formats, our SDAG learns to be aware of the degradation effect on semantic levels in a relative comparison manner. Moreover, our SDAG is general and can be plugged into the training paradigm of the existing ULLIE methods. Extensive experiments demonstrate its effectiveness for improving the ULLIE approaches on the downstream recognition tasks while maintaining a competitive visual quality. Code will be available at https://github.com/zheng980629/SDAG.

Hybrid Pixel-Unshuffled Network for Lightweight Image Super-resolution

Convolutional neural network (CNN) has achieved great success on image super-resolution (SR). However, most deep CNN-based SR models take massive computations to obtain high performance. Downsampling features for multi-resolution fusion is an efficient and effective way to improve the performance of visual recognition. Still, it is counter-intuitive in the SR task, which needs to project a low-resolution input to high-resolution. In this paper, we propose a novel Hybrid Pixel-Unshuffled Network (HPUN) by introducing an efficient and effective downsampling module into the SR task. The network contains pixel-unshuffled downsampling and Self-Residual Depthwise Separable Convolutions. Specifically, we utilize pixel-unshuffle operation to downsample the input features and use grouped convolution to reduce the channels. Besides, we enhance the depthwise convolution's performance by adding the input feature to its output. The comparison findings demonstrate that, with fewer parameters and computational costs, our HPUN achieves and surpasses the state-of-the-art performance on SISR. All results are provided in the github https://github.com/Sun1992/HPUN.

No increased prevalence of prosopagnosia in aphantasia: Visual recognition deficits are small and not restricted to faces.

Aphantasia and prosopagnosia are both rare conditions with impairments in visual cognition. While prosopagnosia refers to a face recognition deficit, aphantasics exhibit a lack of mental imagery. Current object recognition theories propose an interplay of perception and mental representations, making an association between recognition performance and visual imagery plausible. While the literature assumes a link between aphantasia and prosopagnosia, other impairments in aphantasia have been shown to be rather global. Therefore, we assumed that aphantasics do not solely exhibit impairments in face recognition but rather in general visual recognition performance, probably moderated by stimulus complexity. To test this hypothesis, 65 aphantasics were compared to 55 controls in a face recognition task, the Cambridge Face Memory Test, and a corresponding object recognition task, the Cambridge Car Memory Test. In both tasks, aphantasics performed worse than controls, indicating mild recognition deficits without face-specificity. Additional correlations between imagery vividness and performance in both tasks were found, suggesting that visual imagery influences visual recognition not only in imagery extremes. Stimulus complexity produced the expected moderation effect but only for the whole imagery-spectrum and only with face stimuli. Overall, the results imply that aphantasia is linked to a general but mild deficit in visual recognition.

Response assignment influences visual recognition.

This study investigated whether two different neural systems influenced performance in an immediate visual recognition, i.e. visual same/different task. An observer had to respond rapidly whether a test consonant had just appeared in the study string by pressing one of two response keys, labeled same and different. When the same response was assigned to the response key on the right, there was no effect of study-string position on target response time (RT), indicating that the test item was not compared with the study string. When the different response was assigned to the response key on the right, same RT was an increasing function of the left-to-right position of a target in the study string and different RT was slower than same RT, indicating that during test the study string was compared with the test item. Functional magnetic resonance imaging confirmed that the caudate and left hippocampus were more active when the different response was assigned to the right key but the right hippocampus was more active when the same response was assigned to the right key. Therefore, two different computational processes are performed by two different brain systems depending on whether the same or different response is assigned to the right response key.

Application of SWMM model and image recognition in virtual simulation of urban architectural landscape

As the country prioritizes urban water supply issues and water environment, a series of policies have been formulated to manage urban water supply and environmental protection, such as sponge city construction. This article first selects the adaptability study of a city's sponge architecture based on the SWMM model developed by the US Environmental Protection Agency, the overall concept of a city's planning and construction, and watershed water. In today's society, visual recognition technology has been widely used and plays an important role in all aspects. Good recognition technology is the key, and it is of great significance to how to improve the recognition level and recognition speed, which is directly related to the actual performance and safety of visual recognition. In addition, this article focuses on the current structural design and design inefficiency, one-sidedness, and the inability to protect the environment. It also reflects the planning concept for the environment (including people), as well as the process system of building and landscape development, and a brief summary of the virtual simulation technology of the structure. This article selects Sketch Up and 3dsMax for various comparisons for software modeling and design, and combines VR-Platform as a development platform to design and implement virtual 3D vision planning functions. When designing and creating this virtual simulation system, it will mainly focus on four-dimensional landscape, three-dimensional landscape modeling, landscape design and three-dimensional landscape planning.

A deep learning loss based on additive cosine margin: Application to fashion style and face recognition

Recently, loss functions based on angular spans improved the performance of deep visual recognition. These losses converted Euclidean cross entropy to angular cross entropy loss. Fashion style recognition deals with the problem of assigning a person’s outfit to a fashion style category. Due to the high similarity between different clothing items and the use of softmax-based loss functions, many of the current methods that address this problem show relatively poor performance and cannot guarantee sufficient inter-class margins in the fashion domain. In this work, we propose an end-to-end method for deep visual recognition by combining a standard CNN architecture with a novel loss function, which we call Additive Cosine Margin Loss (ACML). The proposed function not only projects feature vectors of different classes into different regions of the embedding, but also enforces compactness of the projections within each class. Our experiments were conducted on two public and well-known fashion style recognition datasets FashionStyle14 and HipsterWars, and on the face verification and identification datasets LFW, YTF, and MegaFace. These experiments demonstrate the superiority of the proposed loss function over: (1) existing angular margin-based loss functions (2) state-of-the-art methods for clothing style recognition as well as face analysis tasks.