Earth observation is fundamental for a range of human activities including flood response as it offers vital information to decision makers. Semantic segmentation plays a key role in mapping the raw hyper-spectral data coming from the satellites into a human understandable form assigning class labels to each pixel. In this paper, we introduce a prototype-based interpretable deep semantic segmentation (IDSS) method, which is highly accurate as well as interpretable. Its parameters are in orders of magnitude less than the number of parameters used by deep networks such as U-Net and are clearly interpretable by humans. The proposed here IDSS offers a transparent structure that allows users to inspect and audit the algorithm's decision. Results have demonstrated that IDSS could surpass other algorithms, including U-Net, in terms of IoU (Intersection over Union) total water and Recall total water. We used WorldFloods data set for our experiments and plan to use the semantic segmentation results combined with masks for permanent water to detect flood events.
Amodal perception requires inferring the full shape of an object that is partially occluded. This task is particularly challenging on two levels: (1) it requires more information than what is contained in the instant retina or imaging sensor, (2) it is difficult to obtain enough well-annotated amodal labels for supervision. To this end, this paper develops a new framework of Self-supervised amodal Video object segmentation (SaVos). Our method efficiently leverages the visual information of video temporal sequences to infer the amodal mask of objects. The key intuition is that the occluded part of an object can be explained away if that part is visible in other frames, possibly deformed as long as the deformation can be reasonably learned. Accordingly, we derive a novel self-supervised learning paradigm that efficiently utilizes the visible object parts as the supervision to guide the training on videos. In addition to learning type prior to complete masks for known types, SaVos also learns the spatiotemporal prior, which is also useful for the amodal task and could generalize to unseen types. The proposed framework achieves the state-of-the-art performance on the synthetic amodal segmentation benchmark FISHBOWL and the real world benchmark KINS-Video-Car. Further, it lends itself well to being transferred to novel distributions using test-time adaptation, outperforming existing models even after the transfer to a new distribution.
Airport runway segmentation can effectively reduce the accident rate during the landing phase, which has the largest risk of flight accidents. With the rapid development of deep learning (DL), related methods achieve good performance on segmentation tasks and can be well adapted to complex scenes. However, the lack of large-scale, publicly available datasets in this field makes the development of methods based on DL difficult. Therefore, we propose a benchmark for airport runway segmentation, named BARS. Additionally, a semiautomatic annotation pipeline is designed to reduce the annotation workload. BARS has the largest dataset with the richest categories and the only instance annotation in the field. The dataset, which was collected using the X-Plane simulation platform, contains 10,256 images and 30,201 instances with three categories. We evaluate eleven representative instance segmentation methods on BARS and analyze their performance. Based on the characteristic of an airport runway with a regular shape, we propose a plug-and-play smoothing postprocessing module (SPM) and a contour point constraint loss (CPCL) function to smooth segmentation results for mask-based and contour-based methods, respectively. Furthermore, a novel evaluation metric named average smoothness (AS) is developed to measure smoothness. The experiments show that existing instance segmentation methods can achieve prediction results with good performance on BARS. SPM and CPCL can effectively enhance the AS metric while modestly improving accuracy. Our work will be available at https://github.com/c-wenhui/BARS.
Few-shot segmentation aims to segment images containing objects from previously unseen classes using only a few annotated samples. Most current methods focus on using object information extracted, with the aid of human annotations, from support images to identify the same objects in new query images. However, background information can also be useful to distinguish objects from their surroundings. Hence, some previous methods also extract background information from the support images. In this paper, we argue that such information is of limited utility, as the background in different images can vary widely. To overcome this issue, we propose CobNet which utilises information about the background that is extracted from the query images without annotations of those images. Experiments show that our method achieves a mean Intersection-over-Union score of 61.4% and 37.8% for 1-shot segmentation on PASCAL-5i and COCO-20i respectively, outperforming previous methods. It is also shown to produce state-of-the-art performances of 53.7% for weakly-supervised few-shot segmentation, where no annotations are provided for the support images.
The segmentation of histopathological whole slide images into tumourous and non-tumourous types of tissue is a challenging task that requires the consideration of both local and global spatial contexts to classify tumourous regions precisely. The identification of subtypes of tumour tissue complicates the issue as the sharpness of separation decreases and the pathologist's reasoning is even more guided by spatial context. However, the identification of detailed types of tissue is crucial for providing personalized cancer therapies. Due to the high resolution of whole slide images, existing semantic segmentation methods, restricted to isolated image sections, are incapable of processing context information beyond. To take a step towards better context comprehension, we propose a patch neighbour attention mechanism to query the neighbouring tissue context from a patch embedding memory bank and infuse context embeddings into bottleneck hidden feature maps. Our memory attention framework (MAF) mimics a pathologist's annotation procedure -- zooming out and considering surrounding tissue context. The framework can be integrated into any encoder-decoder segmentation method. We evaluate the MAF on a public breast cancer and an internal kidney cancer data set using famous segmentation models (U-Net, DeeplabV3) and demonstrate the superiority over other context-integrating algorithms -- achieving a substantial improvement of up to $17\%$ on Dice score. The code is publicly available at: https://github.com/tio-ikim/valuing-vicinity
Unsupervised image segmentation is an important task in many real-world scenarios where labelled data is of scarce availability. In this paper we propose a novel approach that harnesses recent advances in unsupervised learning using a combination of Mutual Information Maximization (MIM), Neural Superpixel Segmentation and Graph Neural Networks (GNNs) in an end-to-end manner, an approach that has not been explored yet. We take advantage of the compact representation of superpixels and combine it with GNNs in order to learn strong and semantically meaningful representations of images. Specifically, we show that our GNN based approach allows to model interactions between distant pixels in the image and serves as a strong prior to existing CNNs for an improved accuracy. Our experiments reveal both the qualitative and quantitative advantages of our approach compared to current state-of-the-art methods over four popular datasets.
Weed and crop segmentation is becoming an increasingly integral part of precision farming that leverages the current computer vision and deep learning technologies. Research has been extensively carried out based on images captured with a camera from various platforms. Unmanned aerial vehicles (UAVs) and ground-based vehicles including agricultural robots are the two popular platforms for data collection in fields. They all contribute to site-specific weed management (SSWM) to maintain crop yield. Currently, the data from these two platforms is processed separately, though sharing the same semantic objects (weed and crop). In our paper, we have developed a deep convolutional network that enables to predict both field and aerial images from UAVs for weed segmentation and mapping with only field images provided in the training phase. The network learning process is visualized by feature maps at shallow and deep layers. The results show that the mean intersection of union (IOU) values of the segmentation for the crop (maize), weeds, and soil background in the developed model for the field dataset are 0.744, 0.577, 0.979, respectively, and the performance of aerial images from an UAV with the same model, the IOU values of the segmentation for the crop (maize), weeds and soil background are 0.596, 0.407, and 0.875, respectively. To estimate the effect on the use of plant protection agents, we quantify the relationship between herbicide spraying saving rate and grid size (spraying resolution) based on the predicted weed map. The spraying saving rate is up to 90% when the spraying resolution is at 1.78 x 1.78 cm2. The study shows that the developed deep convolutional neural network could be used to classify weeds from both field and aerial images and delivers satisfactory results.
Discontinuity in the delineation of peripheral bronchioles hinders the potential clinical application of automated airway segmentation models. Moreover, the deployment of such models is limited by the data heterogeneity across different centres, and pathological abnormalities also make achieving accurate robust segmentation in distal small airways difficult. Meanwhile, the diagnosis and prognosis of lung diseases often rely on evaluating structural changes in those anatomical regions. To address this gap, this paper presents a patch-scale adversarial-based refinement network that takes in preliminary segmentation along with original CT images and outputs a refined mask of the airway structure. The method is validated on three different datasets encompassing healthy cases, cases with cystic fibrosis and cases with COVID-19. The results are quantitatively evaluated by seven metrics and achieved more than a 15% rise in detected length ratio and detected branch ratio, showing promising performance compared to previously proposed models. The visual illustration also proves our refinement guided by a patch-scale discriminator and centreline objective functions is effective in detecting discontinuities and missing bronchioles. Furthermore, the generalizability of our refinement pipeline is tested on three previous models and improves their segmentation completeness significantly.
Left-ventricular ejection fraction (LVEF) is an important indicator of heart failure. Existing methods for LVEF estimation from video require large amounts of annotated data to achieve high performance, e.g. using 10,030 labeled echocardiogram videos to achieve mean absolute error (MAE) of 4.10. Labeling these videos is time-consuming however and limits potential downstream applications to other heart diseases. This paper presents the first semi-supervised approach for LVEF prediction. Unlike general video prediction tasks, LVEF prediction is specifically related to changes in the left ventricle (LV) in echocardiogram videos. By incorporating knowledge learned from predicting LV segmentations into LVEF regression, we can provide additional context to the model for better predictions. To this end, we propose a novel Cyclical Self-Supervision (CSS) method for learning video-based LV segmentation, which is motivated by the observation that the heartbeat is a cyclical process with temporal repetition. Prediction masks from our segmentation model can then be used as additional input for LVEF regression to provide spatial context for the LV region. We also introduce teacher-student distillation to distill the information from LV segmentation masks into an end-to-end LVEF regression model that only requires video inputs. Results show our method outperforms alternative semi-supervised methods and can achieve MAE of 4.17, which is competitive with state-of-the-art supervised performance, using half the number of labels. Validation on an external dataset also shows improved generalization ability from using our method. Our code is available at https://github.com/xmed-lab/CSS-SemiVideo.
Training a large-scale deep neural network in a large-scale dataset is challenging and time-consuming. The recent breakthrough of large-batch optimization is a promising way to tackle this challenge. However, although the current advanced algorithms such as LARS and LAMB succeed in classification models, the complicated pipelines of dense visual predictions such as object detection and segmentation still suffer from the heavy performance drop in the large-batch training regime. To address this challenge, we propose a simple yet effective algorithm, named Adaptive Gradient Variance Modulator (AGVM), which can train dense visual predictors with very large batch size, enabling several benefits more appealing than prior arts. Firstly, AGVM can align the gradient variances between different modules in the dense visual predictors, such as backbone, feature pyramid network (FPN), detection, and segmentation heads. We show that training with a large batch size can fail with the gradient variances misaligned among them, which is a phenomenon primarily overlooked in previous work. Secondly, AGVM is a plug-and-play module that generalizes well to many different architectures (e.g., CNNs and Transformers) and different tasks (e.g., object detection, instance segmentation, semantic segmentation, and panoptic segmentation). It is also compatible with different optimizers (e.g., SGD and AdamW). Thirdly, a theoretical analysis of AGVM is provided. Extensive experiments on the COCO and ADE20K datasets demonstrate the superiority of AGVM. For example, it can train Faster R-CNN+ResNet50 in 4 minutes without losing performance. AGVM enables training an object detector with one billion parameters in just 3.5 hours, reducing the training time by 20.9x, whilst achieving 62.2 mAP on COCO. The deliverables are released at https://github.com/Sense-X/AGVM.
Click-based interactive image segmentation aims at extracting objects with a limited user clicking. A hierarchical backbone is the de-facto architecture for current methods. Recently, the plain, non-hierarchical Vision Transformer (ViT) has emerged as a competitive backbone for dense prediction tasks. This design allows the original ViT to be a foundation model that can be finetuned for downstream tasks without redesigning a hierarchical backbone for pretraining. Although this design is simple and has been proven effective, it has not yet been explored for interactive image segmentation. To fill this gap, we propose SimpleClick, the first interactive segmentation method that leverages a plain backbone. Based on the plain backbone, we introduce a symmetric patch embedding layer that encodes clicks into the backbone with minor modifications to the backbone itself. With the plain backbone pretrained as a masked autoencoder (MAE), SimpleClick achieves state-of-the-art performance. Remarkably, our method achieves 4.15 NoC@90 on SBD, improving 21.8% over the previous best result. Extensive evaluation on medical images demonstrates the generalizability of our method. We further develop an extremely tiny ViT backbone for SimpleClick and provide a detailed computational analysis, highlighting its suitability as a practical annotation tool.
Interactive image segmentation aims at segmenting a target region through a way of human-computer interaction. Recent works based on deep learning have achieved excellent performance, while most of them focus on improving the accuracy of the training set and ignore potential improvement on the test set. In the inference phase, they tend to have a good performance on similar domains to the training set, and lack adaptability to domain shift, so they require more user efforts to obtain satisfactory results. In this work, we propose RAIS, a robust and accurate architecture for interactive segmentation with continuous learning, where the model can learn from both train and test data sets. For efficient learning on the test set, we propose a novel optimization strategy to update global and local parameters with a basic segmentation module and adaptation module, respectively. Moreover, we perform extensive experiments on several benchmarks that show our method can handle data distribution shifts and achieves SOTA performance compared with recent interactive segmentation methods. Besides, our method also shows its robustness in the datasets of remote sensing and medical imaging where the data domains are completely different between training and testing.
Colorectal cancer (CRC) is among the top three malignant tumor types in terms of morbidity and mortality. Histopathological images are the gold standard for diagnosing colon cancer. Cellular nuclei instance segmentation and classification, and nuclear component regression tasks can aid in the analysis of the tumor microenvironment in colon tissue. Traditional methods are still unable to handle both types of tasks end-to-end at the same time, and have poor prediction accuracy and high application costs. This paper proposes a new UNet model for handling nuclei based on the UNet framework, called MGTUNet, which uses Mish, Group normalization and transposed convolution layer to improve the segmentation model, and a ranger optimizer to adjust the SmoothL1Loss values. Secondly, it uses different channels to segment and classify different types of nucleus, ultimately completing the nuclei instance segmentation and classification task, and the nuclei component regression task simultaneously. Finally, we did extensive comparison experiments using eight segmentation models. By comparing the three evaluation metrics and the parameter sizes of the models, MGTUNet obtained 0.6254 on PQ, 0.6359 on mPQ, and 0.8695 on R2. Thus, the experiments demonstrated that MGTUNet is now a state-of-the-art method for quantifying histopathological images of colon cancer.
Automatically segmenting lesion area in breast ultrasound (BUS) images is a challenging one due to its noise, speckle and artifacts. Edge-map of BUS images also does not help because in most cases the edge-map gives no information whatsoever. Almost all segmentation technique takes the edge-map of the image as its first step, though there are a few algorithms that try to avoid edge-maps as well. Improving the edge-map of breast ultrasound images theoretically improves the chances of automatic segmentation to be more precise. In this paper, we propose a semi-automatic technique of histogram splitting using two pointers. Here the user only has to select two initially guessed points denoting a circle on the region of interest (ROI). The method will automatically study the internal histogram and split it using two pointers. The output BUS image has improved edge-map and ultimately the segmentation on it is better compared to regular segmentation using same algorithm and same initialization. Also, we further processed the edge-map to have less edge-pixels to area ratio, improving the homogeneity and the chances of easy segmentation in the future.
Scribble-supervised medical image segmentation tackles the limitation of sparse masks. Conventional approaches alternate between: labeling pseudo-masks and optimizing network parameters. However, such iterative two-stage paradigm is unwieldy and could be trapped in poor local optima since the networks undesirably regress to the erroneous pseudo-masks. To address these issues, we propose a non-iterative method where a stream of varying (pacing) pseudo-masks teach a network via consistency training, named PacingPseudo. Our motivation lies first in a non-iterative process. Interestingly, it can be achieved gracefully by a siamese architecture, wherein a stream of pseudo-masks naturally assimilate a stream of predicted masks during training. Second, we make the consistency training effective with two necessary designs: (i) entropy regularization to obtain high-confidence pseudo-masks for effective teaching; and (ii) distorted augmentations to create discrepancy between the pseudo-mask and predicted-mask streams for consistency regularization. Third, we devise a new memory bank mechanism that provides an extra source of ensemble features to complement scarce labeled pixels. The efficacy of the proposed PacingPseudo is validated on three public medical image datasets, including the segmentation tasks of abdominal multi-organs, cardiac structures, and myocardium. Extensive experiments demonstrate our PacingPseudo improves the baseline by large margins and consistently outcompetes several previous methods. In some cases, our PacingPseudo achieves comparable performance with its fully-supervised counterparts, showing the feasibility of our method for the challenging scribble-supervised segmentation applications. The code and scribble annotations will be publicly available.
Cervical cancer is a public health problem, where the treatment has a better chance of success if detected early. The analysis is a manual process which is subject to a human error, so this paper provides a way to analyze argyrophilic nucleolar organizer regions (AgNOR) stained slide using deep learning approaches. Also, this paper compares models for instance and semantic detection approaches. Our results show that the semantic segmentation using U-Net with ResNet-18 or ResNet-34 as the backbone have similar results, and the best model shows an IoU for nucleus, cluster, and satellites of 0.83, 0.92, and 0.99 respectively. For instance segmentation, the Mask R-CNN using ResNet-50 performs better in the visual inspection and has a 0.61 of the IoU metric. We conclude that the instance segmentation and semantic segmentation models can be used in combination to make a cascade model able to select a nucleus and subsequently segment the nucleus and its respective nucleolar organizer regions (NORs).
Deep Neural Network (DNN) based point cloud semantic segmentation has presented significant achievements on large-scale labeled aerial laser point cloud datasets. However, annotating such large-scaled point clouds is time-consuming. Due to density variations and spatial heterogeneity of the Airborne Laser Scanning (ALS) point clouds, DNNs lack generalization capability and thus lead to unpromising semantic segmentation, as the DNN trained in one region underperform when directly utilized in other regions. However, Self-Supervised Learning (SSL) is a promising way to solve this problem by pre-training a DNN model utilizing unlabeled samples followed by a fine-tuned downstream task involving very limited labels. Hence, this work proposes a hard-negative sample aware self-supervised contrastive learning method to pre-train the model for semantic segmentation. The traditional contrastive learning for point clouds selects the hardest negative samples by solely relying on the distance between the embedded features derived from the learning process, potentially evolving some negative samples from the same classes to reduce the contrastive learning effectiveness. Therefore, we design an AbsPAN (Absolute Positive And Negative samples) strategy based on k-means clustering to filter the possible false-negative samples. Experiments on two typical ALS benchmark datasets demonstrate that the proposed method is more appealing than supervised training schemes without pre-training. Especially when the labels are severely inadequate (10% of the ISPRS training set), the results obtained by the proposed HAVANA method still exceed 94% of the supervised paradigm performance with full training set.
Masked image modeling has demonstrated great potential to eliminate the label-hungry problem of training large-scale vision Transformers, achieving impressive performance on various downstream tasks. In this work, we propose a unified view of masked image modeling after revisiting existing methods. Under the unified view, we introduce a simple yet effective method, termed as MaskDistill, which reconstructs normalized semantic features from teacher models at the masked positions, conditioning on corrupted input images. Experimental results on image classification and semantic segmentation show that MaskDistill achieves comparable or superior performance than state-of-the-art methods. When using the huge vision Transformer and pretraining 300 epochs, MaskDistill obtains 88.3% fine-tuning top-1 accuracy on ImageNet-1k (224 size) and 58.8% semantic segmentation mIoU metric on ADE20k (512 size). The code and pretrained models will be available at https://aka.ms/unimim.
An essential stage in computer aided diagnosis of chest X rays is automated lung segmentation. Due to rib cages and the unique modalities of each persons lungs, it is essential to construct an effective automated lung segmentation model. This paper presents a reliable model for the segmentation of lungs in chest radiographs. Our model overcomes the challenges by learning to ignore unimportant areas in the source Chest Radiograph and emphasize important features for lung segmentation. We evaluate our model on public datasets, Montgomery and Shenzhen. The proposed model has a DICE coefficient of 98.1 percent which demonstrates the reliability of our model.
This paper proposes an efficient iris localization method without using iris segmentation and circle fitting. Conventional iris localization methods first extract iris regions by using semantic segmentation methods such as U-Net. Afterward, the inner and outer iris circles are localized using the traditional circle fitting algorithm. However, this approach requires high-resolution encoder-decoder networks for iris segmentation, so it causes computational costs to be high. In addition, traditional circle fitting tends to be sensitive to noise in input images and fitting parameters, causing the iris recognition performance to be poor. To solve these problems, we propose an iris localization network (ILN), that can directly localize pupil and iris circles with eyelid points from a low-resolution iris image. We also introduce a pupil refinement network (PRN) to improve the accuracy of pupil localization. Experimental results show that the combination of ILN and PRN works in 34.5 ms for one iris image on a CPU, and its localization performance outperforms conventional iris segmentation methods. In addition, generalized evaluation results show that the proposed method has higher robustness for datasets in different domain than other segmentation methods. Furthermore, we also confirm that the proposed ILN and PRN improve the iris recognition accuracy.