Liu Ren

Thang Doan, Xin Li, Sima Behpour et al.

Open World Object Detection (OWOD) is a challenging and realistic task that extends beyond the scope of standard Object Detection task. It involves detecting both known and unknown objects while integrating learned knowledge for future tasks. However, the level of "unknownness" varies significantly depending on the context. For example, a tree is typically considered part of the background in a self-driving scene, but it may be significant in a household context. We argue that this contextual information should already be embedded within the known classes. In other words, there should be a semantic or latent structure relationship between the known and unknown items to be discovered. Motivated by this observation, we propose Hyp-OW, a method that learns and models hierarchical representation of known items through a SuperClass Regularizer. Leveraging this representation allows us to effectively detect unknown objects using a similarity distance-based relabeling module. Extensive experiments on benchmark datasets demonstrate the effectiveness of Hyp-OW, achieving improvement in both known and unknown detection (up to 6 percent). These findings are particularly pronounced in our newly designed benchmark, where a strong hierarchical structure exists between known and unknown objects. Our code can be found at https://github.com/boschresearch/Hyp-OW

Shubhang Bhatnagar, Sharath Gopal, Narendra Ahuja et al.

Gestures form an important medium of communication between humans and machines. An overwhelming majority of existing gesture recognition methods are tailored to a scenario where humans and machines are located very close to each other. This short-distance assumption does not hold true for several types of interactions, for example gesture-based interactions with a floor cleaning robot or with a drone. Methods made for short-distance recognition are unable to perform well on long-distance recognition due to gestures occupying only a small portion of the input data. Their performance is especially worse in resource constrained settings where they are not able to effectively focus their limited compute on the gesturing subject. We propose a novel, accurate and efficient method for the recognition of gestures from longer distances. It uses a dynamic neural network to select features from gesture-containing spatial regions of the input sensor data for further processing. This helps the network focus on features important for gesture recognition while discarding background features early on, thus making it more compute efficient compared to other techniques. We demonstrate the performance of our method on the LD-ConGR long-distance dataset where it outperforms previous state-of-the-art methods on recognition accuracy and compute efficiency.

Yuyan Li, Yuliang Guo, Zhixin Yan et al.

A well-known challenge in applying deep-learning methods to omnidirectional images is spherical distortion. In dense regression tasks such as depth estimation, where structural details are required, using a vanilla CNN layer on the distorted 360 image results in undesired information loss. In this paper, we propose a 360 monocular depth estimation pipeline, OmniFusion, to tackle the spherical distortion issue. Our pipeline transforms a 360 image into less-distorted perspective patches (i.e. tangent images) to obtain patch-wise predictions via CNN, and then merge the patch-wise results for final output. To handle the discrepancy between patch-wise predictions which is a major issue affecting the merging quality, we propose a new framework with the following key components. First, we propose a geometry-aware feature fusion mechanism that combines 3D geometric features with 2D image features to compensate for the patch-wise discrepancy. Second, we employ the self-attention-based transformer architecture to conduct a global aggregation of patch-wise information, which further improves the consistency. Last, we introduce an iterative depth refinement mechanism, to further refine the estimated depth based on the more accurate geometric features. Experiments show that our method greatly mitigates the distortion issue, and achieves state-of-the-art performances on several 360 monocular depth estimation benchmark datasets.

Wenbin He, William Surmeier, Arvind Kumar Shekar et al.

Unsupervised semantic segmentation requires assigning a label to every pixel without any human annotations. Despite recent advances in self-supervised representation learning for individual images, unsupervised semantic segmentation with pixel-level representations is still a challenging task and remains underexplored. In this work, we propose a self-supervised pixel representation learning method for semantic segmentation by using visual concepts (i.e., groups of pixels with semantic meanings, such as parts, objects, and scenes) extracted from images. To guide self-supervised learning, we leverage three types of relationships between pixels and concepts, including the relationships between pixels and local concepts, local and global concepts, as well as the co-occurrence of concepts. We evaluate the learned pixel embeddings and visual concepts on three datasets, including PASCAL VOC 2012, COCO 2017, and DAVIS 2017. Our results show that the proposed method gains consistent and substantial improvements over recent unsupervised semantic segmentation approaches, and also demonstrate that visual concepts can reveal insights into image datasets.

Xin Li, Sima Behpour, Thang Doan et al.

In this study, we investigate the task of data pre-selection, which aims to select instances for labeling from an unlabeled dataset through a single pass, thereby optimizing performance for undefined downstream tasks with a limited annotation budget. Previous approaches to data pre-selection relied solely on visual features extracted from foundation models, such as CLIP and BLIP-2, but largely ignored the powerfulness of text features. In this work, we argue that, with proper design, the joint feature space of both vision and text can yield a better representation for data pre-selection. To this end, we introduce UP-DP, a simple yet effective unsupervised prompt learning approach that adapts vision-language models, like BLIP-2, for data pre-selection. Specifically, with the BLIP-2 parameters frozen, we train text prompts to extract the joint features with improved representation, ensuring a diverse cluster structure that covers the entire dataset. We extensively compare our method with the state-of-the-art using seven benchmark datasets in different settings, achieving up to a performance gain of 20%. Interestingly, the prompts learned from one dataset demonstrate significant generalizability and can be applied directly to enhance the feature extraction of BLIP-2 from other datasets. To the best of our knowledge, UP-DP is the first work to incorporate unsupervised prompt learning in a vision-language model for data pre-selection.

Hengyuan Zhang, David Paz, Yuliang Guo et al.

Autonomous driving for urban and highway driving applications often requires High Definition (HD) maps to generate a navigation plan. Nevertheless, various challenges arise when generating and maintaining HD maps at scale. While recent online mapping methods have started to emerge, their performance especially for longer ranges is limited by heavy occlusion in dynamic environments. With these considerations in mind, our work focuses on leveraging lightweight and scalable priors-Standard Definition (SD) maps-in the development of online vectorized HD map representations. We first examine the integration of prototypical rasterized SD map representations into various online mapping architectures. Furthermore, to identify lightweight strategies, we extend the OpenLane-V2 dataset with OpenStreetMaps and evaluate the benefits of graphical SD map representations. A key finding from designing SD map integration components is that SD map encoders are model agnostic and can be quickly adapted to new architectures that utilize bird's eye view (BEV) encoders. Our results show that making use of SD maps as priors for the online mapping task can significantly speed up convergence and boost the performance of the online centerline perception task by 30% (mAP). Furthermore, we show that the introduction of the SD maps leads to a reduction of the number of parameters in the perception and reasoning task by leveraging SD map graphs while improving the overall performance. Project Page: https://henryzhangzhy.github.io/sdhdmap/.

Abhinav Kumar, Yuliang Guo, Xinyu Huang et al.

Monocular 3D detectors achieve remarkable performance on cars and smaller objects. However, their performance drops on larger objects, leading to fatal accidents. Some attribute the failures to training data scarcity or their receptive field requirements of large objects. In this paper, we highlight this understudied problem of generalization to large objects. We find that modern frontal detectors struggle to generalize to large objects even on nearly balanced datasets. We argue that the cause of failure is the sensitivity of depth regression losses to noise of larger objects. To bridge this gap, we comprehensively investigate regression and dice losses, examining their robustness under varying error levels and object sizes. We mathematically prove that the dice loss leads to superior noise-robustness and model convergence for large objects compared to regression losses for a simplified case. Leveraging our theoretical insights, we propose SeaBird (Segmentation in Bird's View) as the first step towards generalizing to large objects. SeaBird effectively integrates BEV segmentation on foreground objects for 3D detection, with the segmentation head trained with the dice loss. SeaBird achieves SoTA results on the KITTI-360 leaderboard and improves existing detectors on the nuScenes leaderboard, particularly for large objects. Code and models at https://github.com/abhi1kumar/SeaBird

Jinbin Huang, Wenbin He, Liang Gou et al.

Deep learning models are widely used in critical applications, highlighting the need for pre-deployment model understanding and improvement. Visual concept-based methods, while increasingly used for this purpose, face challenges: (1) most concepts lack interpretability, (2) existing methods require model knowledge, often unavailable at run time. Additionally, (3) there lacks a no-code method for post-understanding model improvement. Addressing these, we present InterVLS. The system facilitates model understanding by discovering text-aligned concepts, measuring their influence with model-agnostic linear surrogates. Employing visual analytics, InterVLS offers concept-based explanations and performance insights. It enables users to adjust concept influences to update a model, facilitating no-code model improvement. We evaluate InterVLS in a user study, illustrating its functionality with two scenarios. Results indicates that InterVLS is effective to help users identify influential concepts to a model, gain insights and adjust concept influence to improve the model. We conclude with a discussion based on our study results.

Bilal Alsallakh, Zhixin Yan, Shabnam Ghaffarzadegan et al.

We propose a measure to compute class similarity in large-scale classification based on prediction scores. Such measure has not been formally pro-posed in the literature. We show how visualizing the class similarity matrix can reveal hierarchical structures and relationships that govern the classes. Through examples with various classifiers, we demonstrate how such structures can help in analyzing the classification behavior and in inferring potential corner cases. The source code for one example is available as a notebook at https://github.com/bilalsal/blocks

Yunhao Ge, Hong-Xing Yu, Cheng Zhao et al. · stanford

A major challenge in monocular 3D object detection is the limited diversity and quantity of objects in real datasets. While augmenting real scenes with virtual objects holds promise to improve both the diversity and quantity of the objects, it remains elusive due to the lack of an effective 3D object insertion method in complex real captured scenes. In this work, we study augmenting complex real indoor scenes with virtual objects for monocular 3D object detection. The main challenge is to automatically identify plausible physical properties for virtual assets (e.g., locations, appearances, sizes, etc.) in cluttered real scenes. To address this challenge, we propose a physically plausible indoor 3D object insertion approach to automatically copy virtual objects and paste them into real scenes. The resulting objects in scenes have 3D bounding boxes with plausible physical locations and appearances. In particular, our method first identifies physically feasible locations and poses for the inserted objects to prevent collisions with the existing room layout. Subsequently, it estimates spatially-varying illumination for the insertion location, enabling the immersive blending of the virtual objects into the original scene with plausible appearances and cast shadows. We show that our augmentation method significantly improves existing monocular 3D object models and achieves state-of-the-art performance. For the first time, we demonstrate that a physically plausible 3D object insertion, serving as a generative data augmentation technique, can lead to significant improvements for discriminative downstream tasks such as monocular 3D object detection. Project website: https://gyhandy.github.io/3D-Copy-Paste/

Thang Doan, Sima Behpour, Xin Li et al.

Few-shot Class-Incremental Learning (FSCIL) poses the challenge of retaining prior knowledge while learning from limited new data streams, all without overfitting. The rise of Vision-Language models (VLMs) has unlocked numerous applications, leveraging their existing knowledge to fine-tune on custom data. However, training the whole model is computationally prohibitive, and VLMs while being versatile in general domains still struggle with fine-grained datasets crucial for many applications. We tackle these challenges with two proposed simple modules. The first, Session-Specific Prompts (SSP), enhances the separability of image-text embeddings across sessions. The second, Hyperbolic distance, compresses representations of image-text pairs within the same class while expanding those from different classes, leading to better representations. Experimental results demonstrate an average 10-point increase compared to baselines while requiring at least 8 times fewer trainable parameters. This improvement is further underscored on our three newly introduced fine-grained datasets.

Qinghua Liu, Sam Heshmati, Zheda Mai et al.

Effective analysis of time series data presents significant challenges due to the complex temporal dependencies and cross-channel interactions in multivariate data. Inspired by the way human analysts visually inspect time series to uncover hidden patterns, we ask: can incorporating visual representations enhance automated time-series analysis? Recent advances in multimodal large language models have demonstrated impressive generalization and visual understanding capability, yet their application to time series remains constrained by the modality gap between continuous numerical data and discrete natural language. To bridge this gap, we introduce MLLM4TS, a novel framework that leverages multimodal large language models for general time-series analysis by integrating a dedicated vision branch. Each time-series channel is rendered as a horizontally stacked color-coded line plot in one composite image to capture spatial dependencies across channels, and a temporal-aware visual patch alignment strategy then aligns visual patches with their corresponding time segments. MLLM4TS fuses fine-grained temporal details from the numerical data with global contextual information derived from the visual representation, providing a unified foundation for multimodal time-series analysis. Extensive experiments on standard benchmarks demonstrate the effectiveness of MLLM4TS across both predictive tasks (e.g., classification) and generative tasks (e.g., anomaly detection and forecasting). These results underscore the potential of integrating visual modalities with pretrained language models to achieve robust and generalizable time-series analysis.

Yuliang Guo, Abhinav Kumar, Cheng Zhao et al.

Monocular 3D reconstruction for categorical objects heavily relies on accurately perceiving each object's pose. While gradient-based optimization in a NeRF framework updates the initial pose, this paper highlights that scale-depth ambiguity in monocular object reconstruction causes failures when the initial pose deviates moderately from the true pose. Consequently, existing methods often depend on a third-party 3D object to provide an initial object pose, leading to increased complexity and generalization issues. To address these challenges, we present SUP-NeRF, a Streamlined Unification of object Pose estimation and NeRF-based object reconstruction. SUP-NeRF decouples the object's dimension estimation and pose refinement to resolve the scale-depth ambiguity, and introduces a camera-invariant projected-box representation that generalizes cross different domains. While using a dedicated pose estimator that smoothly integrates into an object-centric NeRF, SUP-NeRF is free from external 3D detectors. SUP-NeRF achieves state-of-the-art results in both reconstruction and pose estimation tasks on the nuScenes dataset. Furthermore, SUP-NeRF exhibits exceptional cross-dataset generalization on the KITTI and Waymo datasets, surpassing prior methods with up to 50\% reduction in rotation and translation error.

Kaiyuan Chen, Kush Hari, Trinity Chung et al.

Cloud robotics enables robots to offload complex computational tasks to cloud servers for performance and ease of management. However, cloud compute can be costly, cloud services can suffer occasional downtime, and connectivity between the robot and cloud can be prone to variations in network Quality-of-Service (QoS). We present FogROS2-FT (Fault Tolerant) to mitigate these issues by introducing a multi-cloud extension that automatically replicates independent stateless robotic services, routes requests to these replicas, and directs the first response back. With replication, robots can still benefit from cloud computations even when a cloud service provider is down or there is low QoS. Additionally, many cloud computing providers offer low-cost spot computing instances that may shutdown unpredictably. Normally, these low-cost instances would be inappropriate for cloud robotics, but the fault tolerance nature of FogROS2-FT allows them to be used reliably. We demonstrate FogROS2-FT fault tolerance capabilities in 3 cloud-robotics scenarios in simulation (visual object detection, semantic segmentation, motion planning) and 1 physical robot experiment (scan-pick-and-place). Running on the same hardware specification, FogROS2-FT achieves motion planning with up to 2.2x cost reduction and up to a 5.53x reduction on 99 Percentile (P99) long-tail latency. FogROS2-FT reduces the P99 long-tail latency of object detection and semantic segmentation by 2.0x and 2.1x, respectively, under network slowdown and resource contention.

Jinbin Huang, Wenbin He, Liang Gou et al.

The emergence of large-scale pre-trained models has heightened their application in various downstream tasks, yet deployment is a challenge in environments with limited computational resources. Knowledge distillation has emerged as a solution in such scenarios, whereby knowledge from large teacher models is transferred into smaller student' models, but this is a non-trivial process that traditionally requires technical expertise in AI/ML. To address these challenges, this paper presents InFiConD, a novel framework that leverages visual concepts to implement the knowledge distillation process and enable subsequent no-code fine-tuning of student models. We develop a novel knowledge distillation pipeline based on extracting text-aligned visual concepts from a concept corpus using multimodal models, and construct highly interpretable linear student models based on visual concepts that mimic a teacher model in a response-based manner. InFiConD's interface allows users to interactively fine-tune the student model by manipulating concept influences directly in the user interface. We validate InFiConD via a robust usage scenario and user study. Our findings indicate that InFiConD's human-in-the-loop and visualization-driven approach enables users to effectively create and analyze student models, understand how knowledge is transferred, and efficiently perform fine-tuning operations. We discuss how this work highlights the potential of interactive and visual methods in making knowledge distillation and subsequent no-code fine-tuning more accessible and adaptable to a wider range of users with domain-specific demands.

Xiaoqi Wang, Wenbin He, Xiwei Xuan et al.

The open-vocabulary image segmentation task involves partitioning images into semantically meaningful segments and classifying them with flexible text-defined categories. The recent vision-based foundation models such as the Segment Anything Model (SAM) have shown superior performance in generating class-agnostic image segments. The main challenge in open-vocabulary image segmentation now lies in accurately classifying these segments into text-defined categories. In this paper, we introduce the Universal Segment Embedding (USE) framework to address this challenge. This framework is comprised of two key components: 1) a data pipeline designed to efficiently curate a large amount of segment-text pairs at various granularities, and 2) a universal segment embedding model that enables precise segment classification into a vast range of text-defined categories. The USE model can not only help open-vocabulary image segmentation but also facilitate other downstream tasks (e.g., querying and ranking). Through comprehensive experimental studies on semantic segmentation and part segmentation benchmarks, we demonstrate that the USE framework outperforms state-of-the-art open-vocabulary segmentation methods.

Wenbin He, Suphanut Jamonnak, Liang Gou et al.

Existing semantic segmentation approaches are often limited by costly pixel-wise annotations and predefined classes. In this work, we present CLIP-S$^4$ that leverages self-supervised pixel representation learning and vision-language models to enable various semantic segmentation tasks (e.g., unsupervised, transfer learning, language-driven segmentation) without any human annotations and unknown class information. We first learn pixel embeddings with pixel-segment contrastive learning from different augmented views of images. To further improve the pixel embeddings and enable language-driven semantic segmentation, we design two types of consistency guided by vision-language models: 1) embedding consistency, aligning our pixel embeddings to the joint feature space of a pre-trained vision-language model, CLIP; and 2) semantic consistency, forcing our model to make the same predictions as CLIP over a set of carefully designed target classes with both known and unknown prototypes. Thus, CLIP-S$^4$ enables a new task of class-free semantic segmentation where no unknown class information is needed during training. As a result, our approach shows consistent and substantial performance improvement over four popular benchmarks compared with the state-of-the-art unsupervised and language-driven semantic segmentation methods. More importantly, our method outperforms these methods on unknown class recognition by a large margin.

Yuyan Li, Zhixin Yan, Ye Duan et al.

Omnidirectional 3D information is essential for a wide range of applications such as Virtual Reality, Autonomous Driving, Robotics, etc. In this paper, we propose a novel, model-agnostic, two-stage pipeline for omnidirectional monocular depth estimation. Our proposed framework PanoDepth takes one 360 image as input, produces one or more synthesized views in the first stage, and feeds the original image and the synthesized images into the subsequent stereo matching stage. In the second stage, we propose a differentiable Spherical Warping Layer to handle omnidirectional stereo geometry efficiently and effectively. By utilizing the explicit stereo-based geometric constraints in the stereo matching stage, PanoDepth can generate dense high-quality depth. We conducted extensive experiments and ablation studies to evaluate PanoDepth with both the full pipeline as well as the individual modules in each stage. Our results show that PanoDepth outperforms the state-of-the-art approaches by a large margin for 360 monocular depth estimation.

Arvind Kumar Shekar, Laureen Lake, Liang Gou et al.

The advent of Convolutional Neural Networks (CNNs) has led to their application in several domains. One noteworthy application is the perception system for autonomous driving that relies on the predictions from CNNs. Practitioners evaluate the generalization ability of such CNNs by calculating various metrics on an independent test dataset. A test dataset is often chosen based on only one precondition, i.e., its elements are not a part of the training data. Such a dataset may contain objects that are both similar and novel w.r.t. the training dataset. Nevertheless, existing works do not reckon the novelty of the test samples and treat them all equally for evaluating generalization. Such novelty-based evaluations are of significance to validate the fitness of a CNN in autonomous driving applications. Hence, we propose a CNN generalization scoring framework that considers novelty of objects in the test dataset. We begin with the representation learning technique to reduce the image data into a low-dimensional space. It is on this space we estimate the novelty of the test samples. Finally, we calculate the generalization score as a combination of the test data prediction performance and novelty. We perform an experimental study of the same for our traffic light detection application. In addition, we systematically visualize the results for an interpretable notion of novelty.

Zhenge Zhao, Panpan Xu, Carlos Scheidegger et al.

The interpretation of deep neural networks (DNNs) has become a key topic as more and more people apply them to solve various problems and making critical decisions. Concept-based explanations have recently become a popular approach for post-hoc interpretation of DNNs. However, identifying human-understandable visual concepts that affect model decisions is a challenging task that is not easily addressed with automatic approaches. We present a novel human-in-the-loop approach to generate user-defined concepts for model interpretation and diagnostics. Central to our proposal is the use of active learning, where human knowledge and feedback are combined to train a concept extractor with very little human labeling effort. We integrate this process into an interactive system, ConceptExtract. Through two case studies, we show how our approach helps analyze model behavior and extract human-friendly concepts for different machine learning tasks and datasets and how to use these concepts to understand the predictions, compare model performance and make suggestions for model refinement. Quantitative experiments show that our active learning approach can accurately extract meaningful visual concepts. More importantly, by identifying visual concepts that negatively affect model performance, we develop the corresponding data augmentation strategy that consistently improves model performance.

Sascha Hornauer, Ke Li, Stella X. Yu et al.

Recent progress in network-based audio event classification has shown the benefit of pre-training models on visual data such as ImageNet. While this process allows knowledge transfer across different domains, training a model on large-scale visual datasets is time consuming. On several audio event classification benchmarks, we show a fast and effective alternative that pre-trains the model unsupervised, only on audio data and yet delivers on-par performance with ImageNet pre-training. Furthermore, we show that our discriminative audio learning can be used to transfer knowledge across audio datasets and optionally include ImageNet pre-training.

Liang Gou, Lincan Zou, Nanxiang Li et al.

Traffic light detection is crucial for environment perception and decision-making in autonomous driving. State-of-the-art detectors are built upon deep Convolutional Neural Networks (CNNs) and have exhibited promising performance. However, one looming concern with CNN based detectors is how to thoroughly evaluate the performance of accuracy and robustness before they can be deployed to autonomous vehicles. In this work, we propose a visual analytics system, VATLD, equipped with a disentangled representation learning and semantic adversarial learning, to assess, understand, and improve the accuracy and robustness of traffic light detectors in autonomous driving applications. The disentangled representation learning extracts data semantics to augment human cognition with human-friendly visual summarization, and the semantic adversarial learning efficiently exposes interpretable robustness risks and enables minimal human interaction for actionable insights. We also demonstrate the effectiveness of various performance improvement strategies derived from actionable insights with our visual analytics system, VATLD, and illustrate some practical implications for safety-critical applications in autonomous driving.

Takanori Fujiwara, Jia-Kai Chou, Shilpika et al.

Dimensionality reduction (DR) methods are commonly used for analyzing and visualizing multidimensional data. However, when data is a live streaming feed, conventional DR methods cannot be directly used because of their computational complexity and inability to preserve the projected data positions at previous time points. In addition, the problem becomes even more challenging when the dynamic data records have a varying number of dimensions as often found in real-world applications. This paper presents an incremental DR solution. We enhance an existing incremental PCA method in several ways to ensure its usability for visualizing streaming multidimensional data. First, we use geometric transformation and animation methods to help preserve a viewer's mental map when visualizing the incremental results. Second, to handle data dimension variants, we use an optimization method to estimate the projected data positions, and also convey the resulting uncertainty in the visualization. We demonstrate the effectiveness of our design with two case studies using real-world datasets.

Bilal Alsallakh, Amin Jourabloo, Mao Ye et al.

Convolutional Neural Networks (CNNs) currently achieve state-of-the-art accuracy in image classification. With a growing number of classes, the accuracy usually drops as the possibilities of confusion increase. Interestingly, the class confusion patterns follow a hierarchical structure over the classes. We present visual-analytics methods to reveal and analyze this hierarchy of similar classes in relation with CNN-internal data. We found that this hierarchy not only dictates the confusion patterns between the classes, it furthermore dictates the learning behavior of CNNs. In particular, the early layers in these networks develop feature detectors that can separate high-level groups of classes quite well, even after a few training epochs. In contrast, the latter layers require substantially more epochs to develop specialized feature detectors that can separate individual classes. We demonstrate how these insights are key to significant improvement in accuracy by designing hierarchy-aware CNNs that accelerate model convergence and alleviate overfitting. We further demonstrate how our methods help in identifying various quality issues in the training data.

Amin Jourabloo, Mao Ye, Xiaoming Liu et al.

Face alignment has witnessed substantial progress in the last decade. One of the recent focuses has been aligning a dense 3D face shape to face images with large head poses. The dominant technology used is based on the cascade of regressors, e.g., CNN, which has shown promising results. Nonetheless, the cascade of CNNs suffers from several drawbacks, e.g., lack of end-to-end training, hand-crafted features and slow training speed. To address these issues, we propose a new layer, named visualization layer, that can be integrated into the CNN architecture and enables joint optimization with different loss functions. Extensive evaluation of the proposed method on multiple datasets demonstrates state-of-the-art accuracy, while reducing the training time by more than half compared to the typical cascade of CNNs. In addition, we compare multiple CNN architectures with the visualization layer to further demonstrate the advantage of its utilization.