Author: Latest Results

Abstract

Current image-text matching methods implicitly align visual-semantic segments within images, and employ cross-modal attention mechanisms to discover fine-grained cross-modal semantic correspondences. Although region-word pairs constitute local matches across modalities, they may lead to inaccurate measurements of relevance when viewed from a global perspective of image-text relationships. Additionally, cross-modal attention mechanisms may introduce redundant or irrelevant region-word alignments, which can reduce retrieval accuracy and limit efficiency. To address these challenges, we propose a Dual perception Attention and local-global Similarity Fusion framework(DASF). Specifically, We combine two types of similarity matching, global and local, to establish a more accurate correspondence between images and text by simultaneously considering global semantics and local details during the matching process. Simultaneously, we integrate dual-perception attention mechanisms to learn the relationship between images and text, utilizing attention polarity to determine the degree of matching and better consider contextual and semantic information, thereby reducing interference from irrelevant regions. Extensive experiments on two benchmark datasets, Flickr30K and MSCOCO, demonstrate the superior effectiveness of our DASF, achieving state-of-the-art performance.

Abstract

Based on a qualitative study of Zimbabwean migrants based in South Africa, who regularly remitted goods and money to Zimbabwe between 2010 and 2020, this paper suggests that at a local level, remittances alleviated poverty with very limited if any transformation of the political economy at the national level. Such remittances promoted consumerism without sustainable investment that can structurally transform the economy. In addition, the dependence on remittances entrenches the culture of migration at the local level, which also contributes to or promotes ethno-tribal fissiparity. In rethinking diaspora remittances in the post-Mugabe era, it is advanced that the seemingly intractable economic and political quagmire in Zimbabwe must be resolved to inspire confidence in the diaspora to pull remittances together for a national socio-economic cause and not local-level band-aid accomplishments which remittances currently do.

Abstract

The aim of the paper is to introduce appropriate tools for quantifying the changes occurred at edge level when analyzing land cover changes. The research is conducted in two directions. Firstly, the binary change index at edge level is introduced and its relationship to the already existing binary change edge (computed at pixel level) is studied. It is proved that a certain inequality between the values of the two indices holds. The theoretical part is completed with real-world examples. The second direction investigated in the paper is the possibility of applying Markov type models, a basic tool in land cover change analysis, in quantifying the changes occurred at edge level. This combined approach is illustrated by a case study and it is shown that it provides complementary information related to the spatial extension of the land cover types and to their distribution. Thus, this type of analysis can improve the methodology of measuring the extent and the implications of land cover changes.

Abstract

Purpose

Patients with post prostatectomy incontinence (PPI) seem to have different needs. Therefore, device post-operative readjustability may be a beneficial feature in PPI management, even though it lacks study support. The purpose of this study is to describe our surgical technique for male sling (MS) implantation, assess outcomes, and the impact of readjustability.

Methods

We performed a retrospective analysis of 89 consecutive patients who underwent PPI correction with MS Argus-T™ (Promedon, Córdoba, Argentina) from 2009 to 2021. The median follow-up was 48 months (12–120). Data were collected in a dedicated database. Perioperative variables were assessed. A descriptive statistical analysis was performed. Clinical and urodynamic variables were correlated with the need for readjustments and success.

Results

In this cohort, objective success was achieved in 80.5% of the patients (65.9% cured and 14.6% improved). A total of 85.4% of the patients met the criteria for subjective success (74.4% cured and 11% improved). For the subgroup of patients who received previous treatment for urethral stricture (US), 79% achieved objective success (63.2% cured, 15.8% improved), and 84.2% achieved subjective success (78.9% cured, 5.3% improved). For the subgroup of patients who received previous radiotherapy (RT) before sling surgery, 68.7% achieved objective success (37.5% cured, 31.2% improved), and 75% achieved subjective success (37.5% cured, 37.5% improved). Procedures for device readjustment were necessary for 27.7% of patients in the total study population. RT and previous US treatment were predictive factors for the need of readjustment, with rates of 66.7% and 61.1% (OR: 8.46; CI: 2.46–29.00; p = 0.001/OR: 6.41; CI: 2.05–20.03; p = 0.001, respectively).

Conclusions

MS adjustability improved success rates, especially among irradiated patients and those with previous US. RT was an adverse predictor of total continence status even after readjustments.

Abstract

Accurately determining whether nodules on CT images of the lung are benign or malignant plays an important role in the early diagnosis and treatment of tumors. In this study, the classification and segmentation of benign and malignant nodules on CT images of the lung were performed using deep learning models. A new approach, C+EffxNet, is used for classification. With this approach, the features are extracted from CT images and then classified with different classifiers. In other phases of the study, a segmentation between benign and malignant was performed and, for the first time, a comparison of nodes was made during segmentation. The deep learning models InceptionV3, DenseNet121, and SeResNet101 were used as backbone models for feature extraction in the segmentation phase. In the classification phase, an accuracy of 0.9798, a precision of 0.9802, a recognition of 0.9798, an F1 score of 0.9798, and a kappa value of 0.9690 were achieved. During segmentation, the highest values of 0.8026 Jacard index and 0.8877 Dice coefficient were achieved.

Abstract

Due to the lack of annotations in target bounding boxes, most methods for weakly supervised target detection transform the problem of object detection into a classification problem of candidate regions, making it easy for weakly supervised target detectors to locate significant and highly discriminative local areas of objects. We propose a weak monitoring method that combines attention and erasure mechanisms. The supervised target detection method uses attention maps to search for areas with higher discrimination within candidate regions, and then uses an erasure mechanism to erase the region, forcing the model to enhance its learning of features in areas with weaker discrimination. To improve the positioning ability of the detector, we cascade the weakly supervised target detection network and the fully supervised target detection network, and jointly train the weakly supervised target detection network and the fully supervised target detection network through multi-task learning. Based on the validation trials, the category mean average precision (mAP) and the correct localization (CorLoc) on the two datasets, i.e., VOC2007 and VOC2012, are 55.2% and 53.8%, respectively. In regard to the mAP and CorLoc, this approach significantly outperforms previous approaches, which creates opportunities for additional investigations into weakly supervised target identification algorithms.