Background To compare the dosimetric outcomes of volumetric arc radiotherapy (RapidArc), intensity-modulated radiotherapy (IMRT), and 3D conformal radiotherapy (3DCRT) plans for primary liver ...tumours. Methods 7 patients with localised unresectable hepatocellular carcinoma were included in this study. CT simulation was done with voluntary deep inspiratory breath-holding after administering 80–90 mL of intravenous iodinated contrast. The mean tumour size was 5.6 cm (range 2.1–9.8) and two patients had multiple lesions (range 1–3). All patients were planned for partial liver irradiation for a total dose of 50–66 Gy in 2 Gy per fraction with conformal techniques. 3DCRT, IMRT, and RapidArc plans were generated using Eclipse planning system v.13 and dosimetric analysis was done to evaluate the plan quality and efficiency, including CI, HI, MU delivered, PTV Dmean and Dmax . IMRT plans had 3–5 fields and RapidArc plans had 3–5 arcs. V10 ,V20 ,V30 and V40 of normal liver and Dmean of organs at risk (OAR) and Dmax of spinal cord were also evaluated. Analysis was done using ANOVA and paired t -test with two tailed p < 0.05. Findings All the three techniques had comparable PTV coverage, dose homogeneity, and OAR sparing. IMRT and Rapid Arc had a significantly better conformity index than 3DCRT ( p = 0.03). The high dose areas within the normal liver; V40 and V30 were significantly lower in RapidArc and IMRT plans ( p = 0.03 and p = 0.04, respectively), although no significant differences were noted between IMRT and RapidArc. One patient could not attain the normal liver constraint V33 <33 Gy, which was attained with both IMRT and RapidArc plans. Interpretation RapidArc and IMRT provide better normal liver sparing and conformity than 3DCRT. However, RapidArc was not better than IMRT for liver protection. Further large trials are required to clearly establish the benefits of IMRT and Rapid Arc techniques to treat primary liver tumours.
A fully autonomous wall-painting robot is the focus of this paper's development, expansion, and implementation efforts. Research efforts have not communicated much about interior wall painting. ...Numerous studies have shown that the chemicals used in painting are detrimental to human health, particularly the respiratory and visual systems. In addition, the monotony, time commitment, and physical exertion inherent in the painting process are hard to resist. With the right combination of human and robotic labour, the entire construction process may be optimized, leading to time and labour savings. Furthermore, it would provide an opportunity to lessen or eliminate human exposure to challenging and dangerous conditions, which would address most of the issues, provide protection, and make it happen more often. The study introduces a new wall painting robot dubbed Smart AI based Wall Painting Robot (SAIWPR). To test how well the suggested scheme works, it is cross-validated with the traditional Autonomous Wall Painting Robot (AWPR). The creation of an unsupervised robotic painting system is driven by the wall painting robot. The design goal is to meet the requirements of being simple, lightweight, inexpensive, and having a short painting time. To control the range of motion and navigate the room, ultrasonic sensors are attached to the arm and the mobile base. The arm's movement and the mobile base's trajectory are both managed by a specialized control system.
Video retrieval using natural language queries requires learning semantically meaningful joint embeddings between the text and the audio-visual input. Often, such joint embeddings are learnt using ...pairwise (or triplet) contrastive loss objectives which cannot give enough attention to ‘difficult-to-retrieve’ samples during training. This problem is especially pronounced in data-scarce settings where the data is relatively small (10% of the large scale MSR-VTT) to cover the rather complex audio-visual embedding space. In this context, we propose to compensate for data scarcity by using domain knowledge to augment supervision. Specifically, in addition to the conventional three samples of a triplet (anchor, positive, and negative), we introduce a fourth term - a partial - to define a margin based partial-order loss. The partials are heuristically sampled such that they semantically lie in the overlap zone between the positives and the negatives, thereby resulting in broader embedding coverage. Our proposals consistently outperform the conventional max-margin and triplet losses and improve the state-of-the-art on MSR-VTT and DiDeMO datasets. To further evaluate our method in data-scarce and low-resource setting, we introduce Rudder - a multilingual video-text retrieval dataset that includes audio and textual captions in Marathi, Hindi, Tamil, Kannada, Malayalam and Telugu. We report benchmark results on Rudder while also observing significant gains using the proposed partial order loss, especially when the language specific retrieval models are jointly trained by availing the cross-lingual alignment across the language-specific datasets.1
Video retrieval using natural language queries requires learning semantically meaningful joint embeddings between the text and the audio-visual input. Often, such joint embeddings are learnt using ...pairwise (or triplet) contrastive loss objectives which cannot give enough attention to 'difficult-to-retrieve' samples during training. This problem is especially pronounced in data-scarce settings where the data is relatively small (10% of the large scale MSR-VTT) to cover the rather complex audio-visual embedding space. In this context, we introduce Rudder - a multilingual video-text retrieval dataset that includes audio and textual captions in Marathi, Hindi, Tamil, Kannada, Malayalam and Telugu. Furthermore, we propose to compensate for data scarcity by using domain knowledge to augment supervision. To this end, in addition to the conventional three samples of a triplet (anchor, positive, and negative), we introduce a fourth term - a partial - to define a differential margin based partialorder loss. The partials are heuristically sampled such that they semantically lie in the overlap zone between the positives and the negatives, thereby resulting in broader embedding coverage. Our proposals consistently outperform the conventional max-margin and triplet losses and improve the state-of-the-art on MSR-VTT and DiDeMO datasets. We report benchmark results on Rudder while also observing significant gains using the proposed partial order loss, especially when the language specific retrieval models are jointly trained by availing the cross-lingual alignment across the language-specific datasets.
In this paper, we present a novel approach to the audio-visual video parsing (AVVP) task that demarcates events from a video separately for audio and visual modalities. The proposed parsing approach ...simultaneously detects the temporal boundaries in terms of start and end times of such events. We show how AVVP can benefit from the following techniques geared towards effective cross-modal learning: (i) adversarial training and skip connections (ii) global context aware attention and, (iii) self-supervised pretraining using an audio-video grounding objective to obtain cross-modal audio-video representations. We present extensive experimental evaluations on the Look, Listen, and Parse (LLP) dataset and show that we outperform the state-of-the-art Hybrid Attention Network (HAN) on all five metrics proposed for AVVP. We also present several ablations to validate the effect of pretraining, global attention and adversarial training.