Fully convolutional neural networks for polyp segmentation in colonoscopy

Early diagnosis is essential for the successful treatment of bowel cancers including colorectal cancer (CRC), and capsule endoscopic imaging with robotic actuation can be a valuable diagnostic tool when combined with automated image analysis. We present a deep learning rooted detection and segmentation framework for recognizing lesions in colonoscopy and capsule endoscopy images. We restructure established convolution architectures, such as VGG and ResNets, by converting them into fully-connected convolution networks (FCNs), fine-tune them and study their capabilities for polyp segmentation and detection. We additionally use shape-from-shading (SfS) to recover depth and provide a richer representation of the tissue’s structure in colonoscopy images. Depth is incorporated into our network models as an additional input channel to the RGB information and we demonstrate that the resulting network yields improved performance. Our networks are tested on publicly available datasets and the most accurate segmentation model achieved a mean segmentation interception over union (IU) of 47.78% and 56.95% on the ETIS-Larib and CVC-Colon datasets, respectively. For polyp detection, the top performing models we propose surpass the current state-of-the-art with detection recalls superior to 90% for all datasets tested. To our knowledge, we present the first work to use FCNs for polyp segmentation in addition to proposing a novel combination of SfS and RGB that boosts performance.

<p>Colorectal cancer (CRC) is one of the most prevalent malignant tumors in Colombia and the world. These neoplasms originate in adenomatous lesions or polyps that must be resected to prevent the disease, which can be done with a colonoscopy. It has been reported that during colonoscopy polyps are detected in 40% of men and 30% of women (hyperplastic, adenomatous, serrated, among others), and, on average, 25% of adenomatous polyps (main quality indicator in colonoscopy). However, these lesions are not easy to observe due to the multiplicity of blind spots in the colon and the human error associated with the examination. Objective: to create a computational method for the automatic detection of colorectal polyps using artificial intelligence in recorded videos of real colonoscopy procedures. Methodology: public databases with colorectal polyps and a data collection built in a University Hospital were used. Initially, all the frames of the videos were normalized to reduce the high variability between databases. Subsequently, the polyp detection task is done with a deep learning method using a convolutional neural network. This network is initialized with weights learned on millions of national images from the ImageNet database. The weights of the network are updated using colonoscopy images, following the tuning technique. <strong></strong></p>

Colon cancer is thought about as the third most regularly identified cancer after Brest and lung cancer. Most colon cancers are adenocarcinomas developing from adenomatous polyps, grow on the intima of the colon. The standard procedure for polyp detection is colonoscopy, where the success of the standard colonoscopy depends on the colonoscopist experience and other environmental factors. Nonetheless, throughout colonoscopy procedures, a considerable number (8-37%) of polyps are missed due to human mistakes, and these missed polyps are the prospective reason for colorectal cancer cells. In the last few years, many research groups developed deep learning-based computer-aided (CAD) systems that recommended many techniques for automated polyp detection, localization, and segmentation. Still, accurate polyp detection, segmentation is required to minimize polyp miss out rates. This paper suggested a Super-Resolution Generative Adversarial Network (SRGAN) assisted Encoder-Decoder network for fully automated colon polyp segmentation from colonoscopic images. The proposed deep learning model incorporates the SRGAN in the up-sampling process to achieve more accurate polyp segmentation. We examined our model on the publicly available benchmark datasets CVC-ColonDB and Warwick- QU. The model accomplished a dice score of 0.948 on the CVC-ColonDB dataset, surpassed the recently advanced state-of-the-art (SOTA) techniques. When it is evaluated on the Warwick-QU dataset, it attains a Dice Score of 0.936 on part A and 0.895 on Part B. Our model showed more accurate results for sessile and smaller-sized polyps.

Colorectal cancer, also known as rectal cancer, is one of the most common forms of cancer, and it can be completely cured with early diagnosis. The most effective and objective method of screening and diagnosis is colonoscopy. Polyp segmentation plays a crucial role in the diagnosis and treatment of diseases related to the digestive system, providing doctors with detailed auxiliary boundary information during clinical analysis. To this end, we propose a novel light-weight feature refining and context-guided network (FRCNet) for real-time polyp segmentation. In this method, we first employed the enhanced context-calibrated module to extract the most discriminative features by developing long-range spatial dependence through a context-calibrated operation. This operation is helpful to alleviate the interference of background noise and effectively distinguish the target polyps from the background. Furthermore, we designed the progressive context-aware fusion module to dynamically capture multi-scale polyps by collecting multi-range context information. Finally, the multi-scale pyramid aggregation module was used to learn more representative features, and these features were fused to refine the segmented results. Extensive experiments on the Kvasir, ClinicDB, ColonDB, ETIS, and Endoscene datasets demonstrated the effectiveness of the proposed model. Specifically, FRCNet achieves an mIoU of 84.9% and mDice score of 91.5% on the Kvasir dataset with a model size of only 0.78 M parameters, outperforming state-of-the-art methods. Models and codes are available at the footnote.1

Colorectal cancer remains a leading cause of cancer-related deaths worldwide, with early detection and removal of polyps being critical in preventing disease progression. Automated polyp segmentation, particularly in colonoscopy images, is a challenging task due to the variability in polyp appearance and the low contrast between polyps and surrounding tissues. In this work, we propose an edge-enhanced network (EENet) designed to address these challenges by integrating two novel modules: the covariance edge-enhanced attention (CEEA) and cross-scale edge enhancement (CSEE) modules. The CEEA module leverages covariance-based attention to enhance boundary detection, while the CSEE module bridges multi-scale features to preserve fine-grained edge details. To further improve the accuracy of polyp segmentation, we introduce a hybrid loss function that combines cross-entropy loss with edge-aware loss. Extensive experiments show that the EENet achieves a Dice score of 0.9208 and an IoU of 0.8664 on the Kvasir-SEG dataset, surpassing state-of-the-art models such as Polyp-PVT and PraNet. Furthermore, it records a Dice score of 0.9316 and an IoU of 0.8817 on the CVC-ClinicDB dataset, demonstrating its strong potential for clinical application in polyp segmentation. Ablation studies further validate the contribution of the CEEA and CSEE modules.

Automatic polyp segmentation from colonoscopy images is an essential step in computer aided diagnosis for colorectal cancer. Most of polyp segmentation methods reported in recent years are based on fully supervised deep learning. However, annotation for polyp images by physicians during the diagnosis is time-consuming and costly. In this paper, we present a novel semi-supervised polyp segmentation via collaborative and adversarial learning of focused and dispersive representations learning model, where focused and dispersive extraction module are used to deal with the diversity of location and shape of polyps. In addition, confidence maps produced by a discriminator in an adversarial training framework shows the effectiveness of leveraging unlabeled data and improving the performance of segmentation network. Consistent regularization is further employed to optimize the segmentation networks to strengthen the representation of the outputs of focused and dispersive extraction module. We also propose an auxiliary adversarial learning method to better leverage unlabeled examples to further improve semantic segmentation accuracy. We conduct extensive experiments on two famous polyp datasets: Kvasir-SEG and CVC-Clinic DB. Experimental results demonstrate the effectiveness of the proposed model, consistently outperforming state-of-the-art semi-supervised segmentation models based on adversarial training and even some advanced fully supervised models.

Colorectal cancer has been recognized as the third most prevalent and the third most lethal cancer worldwide. Accurate detection and segmentation of polyps in the colorectal region play a crucial role in facilitating the early diagnosis of cancer. The rapid advancement of deep learning has greatly accelerated the development of automated polyp detection and segmentation techniques. Although the well-established U-Net and its variants have long been regarded as classical architectures in medical image segmentation, the basic U-Net model still exhibits several limitations when applied to complex endoscopic scenarios. These include noise introduced by skip connections, restricted receptive fields for contextual information, and the absence of effective edge-awareness mechanisms. This review systematically elaborates on the recent research progress of polyp segmentation based on U-Net optimization models. According to different structural enhancement strategies, the existing methods can be broadly categorized into four major types: Depth and residual error optimization. for example, ResUNet++ [1] enhances gradient propagation through residual connections. Attention mechanism integration. such as PraNet [2], which employs reverse attention to focus on salient regions. Boundary and multi-scale feature fusion. exemplified by recent models like CaraNet [3] and BUNet [4], which improve the detection accuracy of small objects and enhance boundary precision. Semantic and cross-modal enhancement. represented by TGA-Net [5] and CLIP-Polyp [6], which leverage vision-language pretraining to achieve improved segmentation performance. Our proposed network, MSEANet, integrates the Edge Feature Extraction (EFE), Cross-layer Context Fusion (CCF), and Selective Edge Attention (SEA) modules into a framework and serves as a recent work representative of collaborative optimization. Experimental results indicate the effectiveness of MSEANet in achieving synergic multi-module fusion for polyp segmentation. This review finds that the research trajectory of the U-Net family has been developed from single-module optimization towards multidimensional fusion. Future research will tend to focus on light-weight framework, multimodal semantic guidance, and enhanced interpretability, designing to further improve the clinical applicability of deep segmentation models.

Accurate polyp segmentation from colonoscopy is essential for the early detection of colorectal cancer. However, the variety of polyps manifested in images and the blurry boundary between a polyp and its surrounding mucosa make segmentation challenging. Hence, it is crucial to accurately identify regional boundaries of polyps in colonoscopy. In this paper, we propose Gated Semantic Boundary Network (Gate-SBNet), which is a novel twostream CNN architecture based on an encoder-decoder framework to segment polyps in colonoscopy images. One stream of Gate-SBNet uses a pre-trained ConvNeXt-B model from the image classification task as the semantic encoder to obtain multi-level semantic features from colonoscopy images. Another branch uses the Semantic Boundary Learning Module (SBLM) to learn boundary features based on multi-level semantic features, which process information in parallel with the semantic encoder. By introducing the Gate Convolution Layer (GCLs) into the SBLM module, the semantic information is converted more accurately to boundary information. Therefore, only boundary-related information will be processed by the SBLM. Then, we merge semantic and boundary features as input to an Unet model to obtain the final segmentation result. Our proposed approach was evaluated on five benchmark datasets: Kvasir, CVC-ClinicDB, CVC-ColonDB, CVC-300, and ETIS-LaribPolypDB. Experiments have demonstrated that it is an efficient architecture and capable of making accurate predictions about object boundaries and significantly improving the performance of finding thin and small objects.

BA-Net: Brightness prior guided attention network for colonic polyp segmentation

Fully convolutional neural network (FCN) has been dominating the game of face detection task for a few years with its congenital capability of sliding-window-searching with shared kernels, which boiled down all the redundant calculation, and most recent state-of-the-art methods such as Faster-RCNN, SSD, YOLO and FPN use FCN as their backbone. So here comes one question: Can we find a universal strategy to further accelerate FCN with higher accuracy, so could accelerate all the recent FCN-based methods? To analyze this, we decompose the face searching space into two orthogonal directions, 'scale' and 'spatial'. Only a few coordinates in the space expanded by the two base vectors indicate foreground. So if FCN could ignore most of the other points, the searching space and false alarm should be significantly boiled down. Based on this philosophy, a novel method named scale estimation and spatial attention proposal (S2AP) is proposed to pay attention to some specific scales in image pyramid and valid locations in each scales layer. Furthermore, we adopt a masked-convolution operation based on the attention result to accelerate FCN calculation. Experiments show that FCN-based method RPN can be accelerated by about 4A— with the help of S2AP and masked-FCN and at the same time it can also achieve the state-of-the-art on FDDB, AFW and MALF face detection benchmarks as well.

How Good Is Your Dentist? How Good Is Your Endoscopist? The Quality Imperative

Automated polyp segmentation from colonoscopy images is crucial for colorectal cancer diagnosis. The accuracy of such segmentation, however, is challenged by two main factors. First, the variability in polyps' size, shape, and color, coupled with the scarcity of well-annotated data due to the need for specialized manual annotation, hampers the efficacy of existing deep learning methods. Second, concealed polyps often blend with adjacent intestinal tissues, leading to poor contrast that challenges segmentation models. Recently, diffusion models have been explored and adapted for polyp segmentation tasks. However, the significant domain gap between RGB-colonoscopy images and grayscale segmentation masks, along with the low efficiency of the diffusion generation process, hinders the practical implementation of these models. To mitigate these challenges, we introduce the Highlighted Diffusion Model Plus (HDM+), a two-stage polyp segmentation framework. This framework incorporates the Highlighted Diffusion Model (HDM) to provide explicit semantic guidance, thereby enhancing segmentation accuracy. In the initial stage, the HDM is trained using highlighted ground-truth data, which emphasizes polyp regions while suppressing the background in the images. This approach reduces the domain gap by focusing on the image itself rather than on the segmentation mask. In the subsequent second stage, we employ the highlighted features from the trained HDM's U-Net model as plug-in priors for polyp segmentation, rather than generating highlighted images, thereby increasing efficiency. Extensive experiments conducted on six polyp segmentation benchmarks demonstrate the effectiveness of our approach.

Polyps are the main biomarker to diagnose colorectal cancer. These polyps are protuberant masses that namely appear around intestinal tract. Currently, the automatic polyp detection is a challenging task because the high shape and appearance variability. Additionally, the videos captured during colonoscopies are namely distorted because the clinical procedure, presenting strong camera motions and illumination changes. This work introduces an automatic strategy to detect polyps by using a pixel level characterization of orientation and curvatures, which are coded using a dense Hough transform.Because the complexity of polyp modelling, an additional appearance model was herein implemented, allowing to filter and to find potential polyp regions. The proposed strategy was evaluated over real sequences of ASU-Mayo Clinic Colonoscopy Video dataset, reporting in average 81% of precision on polyp detection task. This strategy is also able to online track polyps in long sequences, requiring only a first frame delineation.

Early detection and assessment of polyps play a crucial role in the prevention and treatment of colorectal cancer (CRC). Polyp segmentation provides an effective solution to assist clinicians in accurately locating and segmenting polyp regions. In the past, people often relied on manually extracted lower-level features such as color, texture, and shape, which often had problems capturing global context and lacked robustness to complex scenarios. With the advent of deep learning, more and more medical image segmentation algorithms based on deep learning networks have emerged, making significant progress in the field. This paper provides a comprehensive review of polyp segmentation algorithms. We first review some traditional algorithms based on manually extracted features and deep segmentation algorithms, and then describe benchmark datasets related to the topic. Specifically, we carry out a comprehensive evaluation of recent deep learning models and results based on polyp size, taking into account the focus of research topics and differences in network structures. Finally, we discuss the challenges of polyp segmentation and future trends in the field.

As the basis of scenes understanding for autonomous driving, lane segmentation is always a challenge due to the various illumination conditions, heavy traffics and richly-textured roads. Because of the heavily biased distribution of lane/non-lane pixels, it is hard to achieve satisfying results by using image segmentation networks such as fully convolution neural networks (FCN). In this paper, we propose a new loss function to tackle the unbalanced data distribution problem. It has shown that the loss function significantly improves the performance of available segmentation networks such as FCN on the lane segmentation task.