Automated Brain Tumour Detection using Deep Learning via Convolution Neural Networks (CNN)

Kumar, Sanjay; Kumar, Naresh; Rishabh,; Kaur, Inderpreet; Keshari, Vivek

doi:10.31782/ijcrr.2021.13234

Cited by 14 publications

(3 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…I. While most of the parameters are the same as the ones in the original SAC [10] implementation, we increased the number of hidden layers units to 1024 to deal with the high-dimensionality of our state space and we increased the training frequency to 10 timesteps to avoid implicit underparametrization [35].…”

Section: B Training Hyperparametersmentioning

confidence: 99%

Fast Object Segmentation Learning with Kernel-based Methods for Robotics

Ceola

Maiettini

Pasquale

et al. 2021

2021 IEEE International Conference on Robotics and Automation (ICRA)

View full text Add to dashboard Cite

Multi-fingered robotic hands could enable robots to perform sophisticated manipulation tasks. However, teaching a robot to grasp objects with an anthropomorphic hand is an arduous problem due to the high dimensionality of state and action spaces. Deep Reinforcement Learning (DRL) offers techniques to design control policies for this kind of problems without explicit environment or hand modeling. However, training these policies with state-of-the-art model-free algorithms is greatly challenging for multi-fingered hands. The main problem is that an efficient exploration of the environment is not possible for such high-dimensional problems, thus causing issues in the initial phases of policy optimization. One possibility to address this is to rely on off-line task demonstrations. However, oftentimes this is incredibly demanding in terms of time and computational resources.In this work, we overcome these requirements and propose the A Grasp Pose is All You Need (G-PAYN) method for the anthropomorphic hand of the iCub humanoid. We develop an approach to automatically collect task demonstrations to initialize the training of the policy. The proposed grasping pipeline starts from a grasp pose generated by an external algorithm, used to initiate the movement. Then a control policy (previously trained with the proposed G-PAYN) is used to reach and grab the object. We deployed the iCub into the MuJoCo simulator and use it to test our approach with objects from the YCB-Video dataset. The results show that G-PAYN outperforms current DRL techniques in the considered setting, in terms of success rate and execution time with respect to the baselines. The code to reproduce the experiments will be released upon acceptance.

show abstract

Section: B Training Hyperparametersmentioning

confidence: 99%

Fast Object Segmentation Learning with Kernel-based Methods for Robotics

Ceola

Maiettini

Pasquale

et al. 2021

2021 IEEE International Conference on Robotics and Automation (ICRA)

View full text Add to dashboard Cite

show abstract

“…While existing literature primarily focuses on enhancing classification methods using advanced ML and DL techniques [26][27][28][29], the perfection of CS in MRI signal acquisition can potentially improve the overall diagnosis model in terms of both acquisition and classification. Therefore, this proposed work aims to perform classification on compressively sensed MRI and compare it with a state-of-the-art methodology for tumor classification, thus providing insights into the scope and future improvements.…”

Section: Related Workmentioning

confidence: 99%

Machine Learning based tumor diagnosis using compressive sensing in MRI images

Mathew,

Stanley,

Jose

2023

Biomed. Phys. Eng. Express

View full text Add to dashboard Cite

Despite the widespread use of Magnetic Resonance Imaging (MRI) analysis for disease diagnosis, processing and analyzing the substantial amount of acquired data may be challenging. Compressive Sensing (CS) offers a promising solution to this problem. MRI diagnosis can be performed faster and more accurately using CS since it requires fewer data for image analysis. A combination of CS with conventional and Deep Learning (DL) models, specifically VGGNet-16, is proposed for categorizing reconstructed MRI images into healthy and unhealthy. The model is properly trained using a dataset containing both normal and tumor images. The method is evaluated using a variety of parameters, including recall, F1-score, accuracy, and precision. Using the VGGNet-16 model, the proposed work achieved a classification accuracy of 98.7%, which is comparable with another state-of-the-art method based on traditionally acquired MRI images. The results indicate that CS may be useful in clinical settings for improving the efficiency and accuracy of MRI-based tumor diagnosis. Furthermore, the approach could be extended to other medical imaging modalities, possibly improving diagnosis accuracy. The study illustrates how CS can enhance medical imaging analysis, particularly in the context of tumor diagnosis using MRI images. It is necessary to conduct further research to investigate the potential applications of CS in other medical imaging contexts.

show abstract

“…The SNL (1) can be overcome also using Fock states and various other exotic non-Gaussian quantum states, see e.g., Refs. [11,[18][19][20][21][22][23][24][25] and the reviews [26,27]. In Ref.…”

Section: Introductionmentioning

confidence: 99%

Sensitivity of Quantum-Enhanced Interferometers

Dariya

Khalili

2023

Symmetry

View full text Add to dashboard Cite

We review various schemes of quantum-enhanced optical interferometers, both linear (SU(2)) and non-linear (SU(1,1)) ones, as well as hybrid SU(2)/SU(1,1) options, using the unified modular approach based on the Quantum Cramèr–Rao bound (QCRB), and taking into account the practical limitations pertinent to all real-world highly-sensitive interferometers. We focus on three important cases defined by the interferometer symmetry: (i) the asymmetric single-arm interferometer; (ii) the symmetric two-arm interferometer with the antisymmetric phase shifts in the arms; and (iii) the symmetric two-arm interferometer with the symmetric phase shifts in the arms. We show that while the optimal regimes for these cases differ significantly, their QCRBs asymptotically correspond to the same squeezing-enhanced shot noise limit (2), which first appeared in the pioneering work by C. Caves in 1981.We show also that in all considered cases the QCRB can be asymptotically saturated by the standard (direct or homodyne) detection schemes.

show abstract

Automated Brain Tumour Detection using Deep Learning via Convolution Neural Networks (CNN)

Cited by 14 publications

References 17 publications

Fast Object Segmentation Learning with Kernel-based Methods for Robotics

Fast Object Segmentation Learning with Kernel-based Methods for Robotics

Machine Learning based tumor diagnosis using compressive sensing in MRI images

Sensitivity of Quantum-Enhanced Interferometers

Contact Info

Product

Resources

About