Search

[Display omitted] [ABSTRACT FROM AUTHOR]

Traditional supervised image captioning methods usually rely on a large number of images and paired captions for training. However, the creation of such datasets necessitates considerable temporal and human resources. Therefore, we propose a new semi-supervised image captioning algorithm to solve this problem. The proposed method uses a generative adversarial network to generate images that match captions, and uses these generated images and captions as new training data. This avoids the error accumulation problem when generating pseudo captions with autoregressive method and the network can directly perform backpropagation. At the same time, in order to ensure the correlation between the generated images and captions, we introduced the CLIP model for constraints. The CLIP model has been pre-trained on a large amount of image–text data, so it shows excellent performance in semantic alignment of images and text. To verify the effectiveness of our method, we validate on MSCOCO offline "Karpathy" test split. Experiment results show that our method can significantly improve the performance of the model when using 1% paired data, with the CIDEr score increasing from 69.5% to 77.7%. This shows that our method can effectively utilize unlabeled data for image caption tasks. • The proposed method is to generate images for captions instead of captions for images. • CLIP is used to constraint generator to associate images with captions. • Parameter updates can be completed through backpropagation. [ABSTRACT FROM AUTHOR]

Intestinal perforation following endoscopic clip placement is an uncommon complication that has been reported in adult patients but, as far as we now, not in pediatric patients. An 8-month-old female infant with a history of metabolic encephalopathy was admitted with a convulsive disorder requiring mechanical ventilation. On the seventh day of hospitalization in the intensive care unit, the patient experienced three bowel movements with dark blood clots. Despite being hemodynamically stable upon physical examination, her hemoglobin level decreased to 8.6 g/dL. Upper endoscopy findings were normal, and exploratory laparoscopy ruled out Meckel's diverticulum or other visible parietal causes of bleeding. Lower gastrointestinal endoscopy revealed ulcers in the rectum, sigmoid, and terminal ileum, where hemostatic clips were placed. No repeated bleeding events occurred. However, 20 days later she presented with fever, abdominal pain, and distension. Abdominal radiography revealed pneumoperitoneum, leading to an emergency exploratory laparotomy. A perforation of the sigmoid colon from one of the clips was found. We resected the short segment of sigmoid colon that contained the perforation and did an end-to-end anastomosis. She recovered well and had no complications. Hemostatic clips placed endoscopically in the colon can lead to perforation even several weeks after placement. [ABSTRACT FROM AUTHOR]

With artificial intelligence and deep learning development, crop disease recognition methods leverage deep networks for automatic feature learning but rely on the volume of training data. Addressing the scarcity of data in agriculture, few-shot learning (FSL) and multi-modal learning have become focal points. However, existing methods are confined to a single modality or insufficiently exploit cross-modal features. To address this, we propose a multi-modal contrastive learning approach integrating images and text to tackle the problem of small-sample recognition. This method combines CLIP multi-modal pre-training with cross-attention, termed ITIMCA. Experimental validation demonstrates the effectiveness of our approach in cassava leaf disease recognition tasks under natural conditions. Experimental results show that the proposed model achieved an accuracy of 78.00%, a precision of 88.48%, a recall rate of 80.00%, and an F1-Score of 79.00% on the cassava leaf disease identification and classification dataset. These results suggest that the proposed network effectively identifies cassava leaf diseases. • Rewriting high-quality cassava leaf description text using LLM models. • Constructing a multimodal dataset using collected cassava leaf images and rewritten text. • Proposing a novel framework ITIMCA that utilizes image-text contrastive learning and cross-attention. • ITIMCA demonstrates significant effectiveness in identifying cassava leaf diseases under limited data and complex natural conditions. [ABSTRACT FROM AUTHOR]

Recent advancements in unsupervised object re-identification have witnessed remarkable progress, which usually focuses on capturing fine-grained semantic information through partitioning or relying on auxiliary networks for optimizing label consistency. However, incorporating extra complex partitioning mechanisms and models leads to non-negligible optimization difficulties, resulting in limited performance gains. To address these problems, this paper presents a Selective Aggregation Network (SANet) to obtain high-quality features and labels for unsupervised object re-identification, which explores primitive fine-grained information of large-scale pre-trained models such as CLIP and designs customized modifications. Specifically, we propose an adaptive selective aggregation module that chooses a set of tokens based on CLIP's attention scores to aggregate discriminative global features. Built upon the representations output by the adaptive selective aggregation module, we design a dynamic weighted clustering algorithm to obtain accurate confidence-weighted pseudo-class centers for contrastive learning. In addition, a dual confidence judgment strategy is introduced to refine and correct the pseudo-labels by assigning three categories of samples through their noise degree. By this means, the proposed SANet enables discriminative feature extraction and clustering refinement for more precise classification without complex architectures such as feature partitioning or auxiliary models. Extensive experiments on existing standard unsupervised object re-identification benchmarks, including Market1501, MSMT17, and Veri776, demonstrate the effectiveness of the proposed SANet method, and SANet achieves state-of-the-art results over other strong competitors. • SANet with feature refinement and label correction for unsupervised object ReID. • Emphasize the discriminative tokens of features for feature refinement. • Subset division and tailored regularization for progressive noise label correction. • Superior SANet performance on benchmarks, outperforming state-of-the-art methods. [ABSTRACT FROM AUTHOR]

Depth estimation has very broad potential in medical image analysis and is important for applications such as augmented reality surgical navigation and preoperative planning. Compared with segmentation tasks that can obtain ground truth through manual annotation, it is difficult to obtain a large number of real values for depth estimation tasks that are limited by hardware conditions in endoscopic environments. To address the challenge, we propose a novel framework that utilizes segmentation tasks to improve encoder performance in a self-supervised depth estimation network. For the first time, we leverage the Contrastive Language-Image Pre-training (CLIP) method to improve the performance of endoscopy segmentation models. Depth estimation networks can also benefit from this training process indirectly. In addition, we design a semantic-guidance loss function to improve the performance. Our proposed method is systematically evaluated on three datasets. Experiments have verified that the proposed framework can assist the network model in obtaining smaller errors. Compared with other state-of-the-art methods, our framework obtains 0.081 and 0.097 on absolute relative error metrics in quantitative evaluations on SCARED and SERV-CT datasets respectively. In qualitative experiments on real surgery datasets, our proposed method also shows more ideal results. The experiments in this study illustrate that our proposed method can alleviate the problem of difficulty in improving network performance due to the lack of real values of depth data. The visual performance of our approach illustrates the application potential in the clinic. Our method helps doctors obtain depth perception and visual cues simultaneously, thereby reducing the difficulty of surgery and the pain of patients. • A framework combining endoscopic depth estimation and segmentation is proposed. • CLIP strategy is applied to endoscopic image segmentation tasks for the first time. • A loss function performs domain smoothing for different physiological structures. [ABSTRACT FROM AUTHOR]

Copyright of Volume! is the property of Editions Melanie Seteun and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Background Indications of clipping (Clip) or coil embolization (Coil) for unruptured cerebral aneurysms (uAN) was not elaborated because prediction of rupture and risk of treatment are difficult. This study aims to determine the risk–benefit analysis of treating uAN by a comprehensive and retrospective investigation of the adverse events and sequelae in patients treated by our Clip/Coil combined units. Methods Clip and Coil were performed in 141 and 80 patients, respectively; Clip for middle cerebral artery AN and Coil for paraclinoid or basilar apex AN. Worsening of modified Rankin scale or mini-mental state examination was defined as major morbidity. Minor morbidity or transient morbidity was defined as other neurologic deficits. Mortality and these morbidities were considered as serious adverse events. Convulsion or events outside the brain were defined as mild adverse events. Results Total mortality and major morbidity were low. Incidence of serious adverse events was not significantly different between the Clip and Coil (17 patients [12.1%] and 6 patients [7.5%]), but the number of total adverse events was significantly different (32 patients [22.7%] in Clip vs. 8 patients [10.0%] in Coil). Because mild morbidities were significantly more frequent in the Clip (20 patients [14.2%]) compared with the Coil (2 patients [2.5%]). Convulsion occurred in 11 (7.8%) patients in the Clip but none in the Coil. Conclusions Our combined unit decreased the occurrence of mortality/major morbidity; however, minor adverse effects were common, especially in the Clip group because of many intrinsic problems of Clip itself. This result suggests further consideration for the treatment modality for uAN. [ABSTRACT FROM AUTHOR]

Collaborative Learning in Practice (CLiP) is one way of addressing an increase in student placement capacity and potentially improving the student learning experience overall. The aim of this article is to report the findings of a CLiP pilot study undertaken in a London hospital maternity ante- and postnatal ward. A qualitative explorative study design employing a descriptive online survey and individual semi-structured interviews to evaluate the experiences. A London based hospital maternity ante- and postnatal ward seven midwifery students from a mix of years (1–3) and six staff (all trained midwives) The survey results and interview data were transcribed and thematically analysed to identify the barriers and enablers for CLiP Three themes emerged from the data: 1. Preparation for the CLiP pilot, 2. Peer-learning and collaboration as support and resource, 3. Independence and trust as drivers for learning. The results are in-line with previously conducted CLiP studies. This study contributes to the findings around the set-up of CLiP in a demanding London maternity ward. It demonstrates that adequate preparation is vital, including the role of the CLiP educator to raise awareness, provide training and to support CLiP midwives. The CLiP hour appears beneficial since it offers protected reflection time. CLiP appeared to increase the clinical confidence of students, particularly more experienced students, through peer learning and independence. Larger-scale research is needed. [ABSTRACT FROM AUTHOR]

Abstract: In the past, the treatment of iatrogenic gastrointestinal perforations was limited to surgical management or to medical observation. Natural Orifice Transluminal Endoscopic Surgery (NOTES) has paved the way towards the development of reliable endoscopic closure techniques, which can be applicable in accidental perforations of the gastrointestinal tract. When endoscopic treatment is feasible, hemoclips are preferred in smaller perforations, while over-the-scope-clips or a combination of hemoclips, endoloops, and glue are used in larger ones. Endoscopic stitching is rarely utilized, and endoscopic stapling has been practically abandoned. The use of self-expandable covered stents can be considered in the esophagus and duodenum. Broad spectrum antibiotics are recommended in most cases. Clinical follow-up in a medico-surgical unit is mandatory and surgical intervention should not be delayed more than 24h if clinical or biological worsening occurs. Imaging with oral contrast medium is advisable before resumption of oral feeding in the case of large perforations. [Copyright &y& Elsevier]

Endoscopic management of nonmalignant leaks, perforations, and ruptures becomes increasingly important with the growing availability of endoscopic therapy. In general, treatment strategies depend on the time of intervention (immediate complication management vs therapy of chronic fistulas and leaks). The approach used to close an acute perforation is adopted from surgical closure techniques, although a serosa-to-serosa adaptation is difficult to achieve, even with new devices. Nonacute leaks—if not treated by stents—are more complex and treatment is based on secondary wound closure after implantation of certain materials and devices. Nevertheless, apart from these general considerations it is important to know that therapy is not standardized and is therefore often based on a patient''s individual factors, availability of respective devices, and knowledge in handling them. [Copyright &y& Elsevier]

Abstract: Background: The “clip and drop” (CD) has been proposed as a useful strategy in the management of severe multifocal necrotizing enterocolitis (NEC). There is little published data on clinical outcomes after this intervention. We report a 2-center experience with this technique. Methods: A retrospective review of infants who underwent CD between 1998 and 2006 at 2 tertiary pediatric surgery centers. Data recorded included intestinal resections, interval between laparotomies, anastomoses at subsequent surgery, time to full feeds, and complications including mortality. Data are reported as median with ranges. Results: Thirteen infants (7 male, 6 female) with a birth weight of 811 (514-2110) g underwent CD of up to 5 bowel segments. In 8 of 9 early survivors, all CD segments were viable. Six patients (46%) were alive at 29 (9-96) months. Survivors underwent 4 (3-4) laparotomies and 4 (2-6) bowel anastomoses and had intestinal continuity restored at 67 (51-162) days. Conclusions: With multiple interventions, half the infants in this high-risk group survived and achieved full enteral feeds. The CD is a valuable technique in a selected group of infants with fulminant NEC. [Copyright &y& Elsevier]

In animal cerebral infarct experiments, the most important aspect is to produce consistent infarct size and localization. In an attempt to improve the conventional middle cerebral artery (MCA) coagulation technique, we developed a new animal model using a microclip to reduce variability in infarct size. Male Sprague–Dawley rats were subjected to right MCA occlusion. The animals were divided into two groups; conventional MCA occlusion group (Group 1; n = 9) and modified clip occlusion group (Group 2; n = 9). In Group 2, the proximal portion of MCA was occluded by applying a small clip just proximal to the olfactory nerve, and the MCA from the clipped position to the position just proximal to the level of the inferior cerebral vein was electrocoagulated using a bipolar diathermy in the same manner as in Group 1. In other words, the only difference between these two groups was the manner of occlusion of the most proximal portion of the MCA. Rats were killed 24 hours after the stroke-inducing surgery, and infarct volume was determined by an image analysis program following staining with 2,3,5-triphenyltetrazolium chloride. The cortical infarct volumes were 51.0 ± 13.8% in Group 1 and 46.3 ± 6.2% in Group 2. The scattering of cortical infarct volume was significantly small in Group 2 (p=0.0176). The differences in scattering of striatal and total infarct volumes did not reach statistical significance. The present results demonstrated that the new MCA occlusion model using a clip significantly reduces the variability in cortical infarct volume, solving the problems of the model using coagulation alone. That permanent MCA occlusion model using a clip is an excellent method that produces more consistent and reproducible infarction. [ABSTRACT FROM AUTHOR]

Endoscopic clip devices are valuable tools for gastrointestinal endoscopists. This review describes in detail the operation of currently available devices and summarizes the evidence for their clinical application. [Copyright &y& Elsevier]

Future developments and innovations in the field of endoscopic hemostasis for upper gastrointestinal bleeding need to be based on an understanding of the efficacy and limitations of current methods and on an understanding of the vascular pathology of bleeding abnormalities. Currently there is a need for more effective hemostasis, especially for bleeding vessels of 1 to 4 mm in diameter which cannot be stopped by injection or thermal methods. There is likely to be an increasing requirement for effective hemostasis at flexible endoscopy in part due to changing disease patterns and also due to an increasing use of flexible endosurgery for example endoscopic mucosal resection (EMR) or endoscopic submucosal dissection (ESD) and transgastric approaches. This article reviews potential improvements in mechanical, thermal, and injection methods that may lead to improvements in flexible endoscopic hemostasis in the upper gastrointestinal tract. [Copyright &y& Elsevier]

Introduction: Blood blister aneurysms (BBAs) are rare aneurysms affecting non-branched points of intracerebral arteries. Due to their small size and fragility, BBAs are prone to rupture, and can be challenging to diagnose and treat. Several treatment options have been suggested yet there is no consensus regarding the best modality to reduce morbidity and mortality.Materials and Methods: A systematic review of the literature was conducted searching for articles discussing the treatment of BBAs. Inclusion criteria included: articles published between January 2010 and August 2020, English language, with each paper including at least 15 patients. Studies included required detailed reporting of patient demographics, treatment, and patient outcomes (including complications, recurrence, neurologic functional status, and mortality).Results and Discussion: A total of 25 studies with 883 patients were included. Most were female (n = 594, 67.3%) and aneurysms were overwhelmingly located in the supraclinoid internal carotid artery (99%). Aneurysms were variable in size and mostly presented with subarachnoid haemorrhage. Endovascular treatment (n = 518, 58.7%) was more common than microsurgery (n = 365, 41.1%) while only 2 patients were managed conservatively. Complications were more common in patients treated microsurgically. Microsurgical procedures had an unfavorable outcome (mRS 4-6, GOS 1-3) rate of 27.8% (n = 100/360) while that of endovascular procedures was 14.7% (n = 70/477). Endovascular procedures had a lower mortality rate than microsurgical interventions (8.4% vs 11%).Conclusion: This review demonstrates that endovascular treatment of blood blister aneurysm has reduced morbidity and mortality when compared with microsurgical treatment. Small sample sizes and substantial study heterogeneity makes strong conclusions difficult. [ABSTRACT FROM AUTHOR]

The objective of this video 1 is to describe the technique to avoid postoperative lymphorrhea after a lumboaortic lymphadenectomy. All procedures were performed at Donostia University Hospital, a tertiary referral and educational center in San Sebastián, Spain. Lumboaortic extra-peritoneal lymphadenectomy was performed for several gynecological malignancies (endometrial and cervical cancer). During the procedure, afferent lymphatic capillaries were identified at the infra-renal aortic level and clipped to avoid retrograde lymphorrhea at this level. Numerous strategies have been described to reduce the likelihood of lymphorea and lymphocele formation. 1 Harmonic scalpel and other sealing advanced devices are not useful to secure lymphatic leakage at this level, although some authors have published a clinical benefit in their use, 2 while clips have been found useful to prevent leakage in other lymphatic locations. 3 The use of harmonic scalpel, biological agents or surgical patch has been ineffective in our experience, but sealing clips and peritonization (marsupialization), 4 once the procedure is concluded, could be an effective approach. Performing simple gestures during lumboaortic lymphadenectomy can help to reduce the appearance of posterior lymphorrhea. [ABSTRACT FROM AUTHOR]

Summary Anal fistulas require surgical treatment to cure the fistula. Fistulotomy is the treatment of choice because of its high success rate, but it also carries a risk of postoperative incontinence. Different methods have been proposed to overcome the need for dividing part or all of the sphincter, so-called “sphincter saving techniques” (SST), such as flap advancement, intrafistular injection of glue, or the insertion of a bio-prosthesis (collagen plug). However, the success rate of SSTs is lower than that of fistulotomy. Ligation of the Intersphincteric Fistula Tract (LIFT) is one of the most recent SSTs. It aims to ligate and transect the fistula by an approach through the intersphincteric space. The methodological quality of most published studies has been only average, but several studies have reported attractive success rates of 70 to 95% without postoperative incontinence. Other techniques such as endo-anal clip or injection of autologous stem cells are still anecdotal. [ABSTRACT FROM AUTHOR]

Summary Anal fistulas require surgical treatment to cure the fistula. Fistulotomy is the treatment of choice because of its high success rate, but it also carries a risk of postoperative incontinence. Different methods have been proposed to overcome the need for dividing part or all of the sphincter, so-called “sphincter saving techniques” (SST), such as flap advancement, intrafistular injection of glue, or the insertion of a bio-prosthesis (collagen plug). However, the success rate of SSTs is lower than that of fistulotomy. Ligation of the Intersphincteric Fistula Tract (LIFT) is one of the most recent SSTs. It aims to ligate and transect the fistula by an approach through the intersphincteric space. The methodological quality of most published studies has been only average, but several studies have reported attractive success rates of 70 to 95% without postoperative incontinence. Other techniques such as endo-anal clip or injection of autologous stem cells are still anecdotal. [ABSTRACT FROM AUTHOR]