No abstract

No abstract

No abstract

No abstract

Accurate relative pose is one of the key components in visual odometry (VO) and simultaneous localization and mapping (SLAM). Recently, the self-supervised learning framework that jointly optimizes the relative pose and target image depth has attracted the attention of the community. Previous works rely on the photometric error generated from depths and poses between adjacent frames, which contains large systematic error under realistic scenes due to reflective surfaces and occlusions. In this paper, we bridge the gap between geometric loss and photometric loss by introducing the matching loss constrained by epipolar geometry in a self-supervised framework. Evaluated on the KITTI dataset, our method outperforms the state-of-the-art unsupervised egomotion estimation methods by a large margin. The code and data are available at https://github.com/hlzz/DeepMatchVO.

No abstract

Neoadjuvant therapy followed by surgery is a standard treatment for locally advanced oesophageal cancer. However, the roles of neoadjuvant chemoradiotherapy and chemotherapy in treating oesophageal cancer remain controversial. In this comprehensive meta-analysis, we examine the efficacy of adding radiotherapy to neoadjuvant chemotherapy for treating oesophageal cancer as reported in qualified randomized controlled trials (RCTs). We conducted a systematic literature search using PubMed, Embase, Cochrane Library databases, Google Scholar and the American Society of Clinical Oncology database to identify relevant studies up to 31 March 2016. Data including the pathological complete response rate, R0 resection rate and 3-year survival rate were extracted and analysed. Five qualified RCTs were included with a total of 709 patients. Meta-analysis showed that neoadjuvant chemoradiotherapy significantly increases the rates of pathological complete response and R0 resection in patients with oesophageal adenocarcinoma or squamous cell carcinoma (SCC). However, we found a significantly increased 3-year survival rate only in oesophageal SCC patients treated with neoadjuvant chemoradiotherapy compared with neoadjuvant chemotherapy (56.8 and 42.8%, respectively); relative risk (RR): 1.31 [95% confidence interval (CI) 1.10-1.58, P = 0.003]. In oesophageal adenocarcinoma patients, no significant survival benefit of neoadjuvant chemoradiotherapy was found compared with neoadjuvant chemotherapy alone (46.3 and 41.0%, respectively; RR: 1.13, 95% CI 0.88-1.45, P = 0.34). Our meta-analysis adds to the evidence showing that neoadjuvant chemoradiotherapy should be the standard preoperative treatment strategy for locally advanced oesophageal SCC. For oesophageal adenocarcinoma, neoadjuvant chemotherapy alone may be the best preoperative treatment strategy to avoid the risk of adverse effects of radiotherapy.

Esophageal squamous cell carcinoma (ESCC) still has a poor prognosis. The prognostic biomarkers of ESCC are not yet well established. Long noncoding RNAs (lncRNAs) have recently been intensively investigated in various cancers including ESCC, and are found to be closely correlated to ESCC. Dysregulated expression of lncRNAs was widely observed in ESCC tumor tissue and was closely related to the tumorigenesis and progression of ESCC. More and more studies have found that lncRNAs were significantly correlated with the prognosis and diagnosis of patients with ESCC. Therefore, all those accumulating evidence indicated that lncRNAs could serve as a prognostic biomarker of ESCC. In this, we summarized the relation between lncRNAs and ESCC as well as the potential biomarker role of lncRNAs in ESCC, especially the prognostic value of lncRNAs. Our current review highlighted the need of further studies to explore the biomarker functions as well as therapeutic values of lncRNAs in ESCC.