Pattern Recognition - Science topic

if possible to submit paper on mobile apps dev?

Great

Thanks for sharing

impact factor journal and what is the cost of paper publications ?

Dear Manuel G Forero ,

25 - 28 November, 2025 Venue: Bogotá, Colombia (Universidad Nacional de Colombia) Site: https://ingenieria.unal.edu.co/ciarp/ Site to submit contributions: https://meteor.springer.com/CIARP2025

Regards,

Shafagat

Muhammad Hamza

Measuring dissimilarities between two image layouts—especially for UI elements like buttons, text boxes, and text—can be effectively approached through a combination of computer vision and layout analysis techniques. One common method involves first performing element detection and segmentation using object detection algorithms such as YOLO, Faster R-CNN, or Detectron2. These models can be trained or fine-tuned to detect specific UI components. Once detected, the bounding boxes of elements (buttons, texts, etc.) can be compared across the two images to assess differences in alignment, positioning, and overlap.

To quantify dissimilarity, you can compute metrics such as Intersection over Union (IoU) for element placement, Euclidean distance between corresponding element centers, and alignment consistency checks (e.g., checking if elements are aligned along common axes). For overlap detection, calculating the area of intersection between text and other elements’ bounding boxes can highlight undesired layout issues.

Tools like Sikulix, Applitools, or Google’s UIAutomator Viewer provide visual diffing capabilities and layout comparisons. More advanced custom pipelines can use OpenCV in combination with OCR engines (like Tesseract) to extract and analyze layout structures. Additionally, recent research in document image analysis and screen similarity learning offers deep-learning-based techniques (e.g., Siamese networks) to learn layout dissimilarities directly. The best approach often depends on whether you need pixel-level comparison or semantic structure awareness.

Mathematical modeling and pattern recognition strengthen children's cognitive skills by activating analytical reasoning early on. These methods train the brain to detect structure in uncertainty—improving problem-solving, memory recall, and attention span through repeated exposure to rules, variations, and abstraction.

In my published research on big data pedagogy, I demonstrated how pattern-based learning models can enhance learner engagement and outcomes by framing complexity through intuitive visual and logical patterns. This framework—although designed for undergraduate analytics—can be scaled down for children using simplified logic structures, gamified rule discovery, and task-based learning aligned with cognitive development stages.

When implemented correctly, these tools simulate early-stage machine learning behavior in humans—encouraging exploration, feedback-based correction, and hierarchical generalization. The key is balancing challenge and clarity, which helps reduce passive screen time while promoting active, critical engagement.

Please read on research on big data pedagogy for more details :-

One of the most powerful ways Generative AI can support traditional Machine Learning is by helping solve the “not enough data” problem, something that comes up a lot in real-world applications.

Take medical imaging as an example. Training a model to detect rare diseases from MRI or CT scans sounds amazing, but the challenge is, those rare cases are exactly that: rare. There's often not enough data to build a reliable model.

That’s where Generative AI, especially models like GANs (Generative Adversarial Networks) and Diffusion Models, comes in as they can actually create high-quality synthetic images that closely resemble real medical scans. These aren’t just random pictures; they follow the same patterns, structures, and styles as real ones.

This is where my thinking stands so far, but I’m sure there’s room for improvement, and I’d love to hear how others are approaching this.

computer science & IT? article charges

I’ve recently published a short research paper titled:

" A Numerical Simulation of Spherical Tank Drainage Using the Modified Euler Method: A Physics-Based Approach "

It presents a simplified yet insightful Modified Euler’s method for solving differential equations and an physical applications on it.

Feel free to check it out, and if you find it useful, I’d really appreciate your recommendation on ResearchGate.

[https://www.researchgate.net/publication/390947672_A_Numerical_Simulation_of_Spherical_Tank_Drainage_Using_the_Modified_Euler_Method_A_Physics-Based_Approach ]

The IEEE 4th International Conference on Digital Society and Intelligent Systems (DSInS 2024) is calling for papers, taking place in two locations: Sydney, Australia, from November 20-22, and Zhengzhou, China, from November 22-24, 2024. Organized by the University of Technology Sydney and Southwest Jiaotong University, with collaboration from several institutions, the conference will focus on advancements in intelligent systems and their applications within digital societies. Topics include but are not limited to machine learning, natural language processing, smart manufacturing, and digital transformation in various sectors. All submissions will undergo a rigorous peer review process, and accepted papers will be published by IEEE and indexed in IEEE Xplore, EI, and Scopus. For more information and to submit papers, please visit the conference website: [DSInS 2024](https://ais.cn/u/m6vUNv). Please note that submissions in the humanities are not accepted.

Md. Faishal Rahaman Final paper submission date is September 30, 2024. For more detail about the conference, please visit the conference website: https://ais.cn/u/MvUV7n

Wishing you every success, International Journal of Complexity in Applied Science and Technology

Thanks for sharing. Wishing you every success in your task.

Thanks for sharing. I wish you every success in your task.

The paper is now available on arxiv: http://arxiv.org/abs/2405.05619. I release the code and dataset publicly: https://paperswithcode.com/paper/rectified-gaussian-kernel-multi-view-k-means

There are several Scopus-indexed journals related to computer science, artificial intelligence, machine learning, Pattern recognition, natural language processing, and Social Media Analytics. without APC or submission fees, and their review process time varies. I have provided the links:

INTERNATIONAL JOURNAL OF INTERACTIVE MULTIMEDIA AND ARTIFICIAL INTELLIGENCE

ISSN: 1989-1660

Vietnam Journal of Computer Science

ISSN (print): 2196-8888 | ISSN (online): 2196-8896

Journal of Applied Research and Technology

ISSN 1665-6423

e-ISSN 2448-6736

Dear Dibyanshu Bhatt ,

Yes, the conference is hybrid format,both online and offline could be accepted.

Submitting your papers to the system is free. Once your paper is accepted, you will need to pay the registration fee. The registration fee could be refer to the website: http://mlise.org/registration

Dear teammates,

I am high experienced in clustering by optimization algorithms such as genetic algorithm, SA, particle swarm optimization algorithm and etc. So, I think I'm skilled to join your group. Please let me know if think so.

Thank you

Qamar Ul Islam :) Is this a ChatGPT-generated answer? It is not useful at all.

"Artificial intelligence (AI) is no longer confined to mere automation; it is now an active participant in the pursuit of knowledge and understanding. However, in the realm of scientific research, the integration of AI marks a significant paradigm shift, ushering in an era where machines and human actively collaborate to formulate research hypotheses and questions. While AI systems have traditionally served as powerful tools for data analysis, their evolution now allows them to go beyond analysis and generate hypotheses, prompting researchers to explore uncharted domains of research..."

Convolutional Neural Networks (CNNs) have been highly successful in various image analysis tasks, including cancer detection. However, traditional CNNs treat all image regions equally when making predictions, which might not be optimal when certain regions contain critical information for cancer detection. To address this, incorporating an attention mechanism into CNNs can significantly improve performance.

Attention mechanisms allow the model to focus on the most informative parts of the image while suppressing less relevant regions. The attention mechanism can be applied to different levels of CNN architectures, such as at the pixel level, spatial level, or channel level. By paying more attention to relevant regions, the CNN with an attention mechanism can enhance the model's ability to detect subtle patterns and features associated with cancerous regions in medical images.

When using CNNs with attention mechanisms for cancer detection, it is crucial to have a sufficiently large dataset with labeled medical images to train the model effectively. Transfer learning with pre-trained models on large-scale image datasets can also be useful to leverage existing knowledge and adapt it to the cancer detection task with a smaller dataset.

Remember that implementing and training deep learning models for cancer detection requires expertise in both deep learning and medical image analysis. Additionally, obtaining annotated medical image datasets and ensuring proper validation and evaluation are essential for developing an accurate and robust cancer detection system. Collaborating with medical professionals and researchers is often necessary to ensure the clinical relevance and accuracy of the developed methods.

To create a neural network with numerical values as input and an image as output, we can use a deep learning library like TensorFlow or PyTorch. In this example, I'll provide you with a simple implementation using TensorFlow and Keras. This implementation will demonstrate how to generate images from random numerical values using a fully connected neural network. Keep in mind that for more complex image generation tasks, you might need a more sophisticated architecture like a Variational Autoencoder (VAE) or a Generative Adversarial Network (GAN).

Before running the code, make sure you have TensorFlow and Keras installed. You can install them using pip:

pip install tensorflow keras

import numpy as np import tensorflow as tf from tensorflow.keras import layers, models # Define the input size for the numerical values input_size = 100 # Define the output image size (e.g., 28x28 grayscale image) output_image_size = (28, 28, 1) # Function to create the generator model def create_generator(): model = models.Sequential() model.add(layers.Dense(256, input_dim=input_size, activation='relu')) model.add(layers.Dense(512, activation='relu')) model.add(layers.Dense(1024, activation='relu')) model.add(layers.Dense(np.prod(output_image_size), activation='sigmoid')) model.add(layers.Reshape(output_image_size)) return model # Function to create random noise as input for the generator def generate_random_noise(batch_size, input_size): return np.random.rand(batch_size, input_size) # Function to create and compile the combined model def create_combined_model(generator, optimizer): generator.trainable = False model = models.Sequential([generator]) model.compile(loss='binary_crossentropy', optimizer=optimizer) return model # Main function def main(): # Generator model generator = create_generator() # Optimizer for the generator optimizer = tf.keras.optimizers.Adam(learning_rate=0.0002, beta_1=0.5) # Combined model combined_model = create_combined_model(generator, optimizer) # Number of epochs and batch size epochs = 20000 batch_size = 128 # Training loop for epoch in range(epochs): # Generate random noise as input for the generator noise = generate_random_noise(batch_size, input_size) # Generate fake images using the generator generated_images = generator.predict(noise) # Your code here: Instead of random noise, you should use your numerical values as input and preprocess them accordingly. # Train the combined model by passing the generated images as inputs and all 1s as the target (since they are fake) d_loss_fake = combined_model.train_on_batch(generated_images, np.ones((batch_size, 1))) # Print the progress if epoch % 100 == 0: print(f"Epoch: {epoch}, Loss: {d_loss_fake}") # Save generated images occasionally if epoch % 1000 == 0: # Your code here: Save the generated images using your preferred method (e.g., matplotlib or OpenCV). pass if __name__ == "__main__": main()

Sure

Pietro Barbetta

OF course you don't (I would guess that -- for most). Read a lot of ME (my writngs, ALL here on researchgate) TO FIND THE WAY to understand.

From a detection perspective, there are some metrics for describing the capacity of accurate edge representation without the need for a reference; these would include measurements of the modulation transfer function (MTF) provided by ISO12233:2014 for example (and more recent versions of ISO12233). However, these are not always appropriate, which is what we in the IEEE P2020 Working Group have been trying to address. There are also Detectivity Metrics, have a look at

Dear Keshav Kumar ,

Look the link, maybe useful.

Regards,

Shafagat

Refer the one in python with seasonality that should resolve the issue https://www.bounteous.com/insights/2020/09/15/forecasting-time-series-model-using-python-part-two/?fbclid=IwAR1Z0V6jS4ruRrpbneHqqzrUA-_fu8nCV_ZXD4GdXSx4UP8WmcEbeCWpbsU

Will you issue certificate

Okay Stay but I think in order for a pattern to exist a theme must recur, therefor it has a characteristic by which it abides, so that is a restriction or physics in a straight jacket as Feynmann calls it. I think it is much like improvising on a musical intsrument, you have to develop your ideas according to a rhythmic cycle. The proton-second is an abstract idea that can be applied to large amounts of particles over macroscopic time periods. Alas the timescale of a second is characteristic of the proton, in this paper it determines its radius.

Dear Philipp Brüggemann

This paper will be useful:

Dear Nisar Ahmed

Done. Good Luck.

Go to ACM computer survey and find out some survey or systematic literature reviews related to your work. You will definitely get some useful insights.

Dear Fikadu Wayesa

Thank you

you can check that link it helps

Busari Yusuf Thank you. I will explore this more to understand the concept of PCA.

Also check https://www.researchgate.net/publication/336804733_Intelligent_fault_diagnosis_system_based_on_big_data

It is depend on the size of input vector, type and size of your data. if the size of data is small, you need features extraction. Remember you need meaningful features!

Dear Sindhuraj Mukherjee,

I suggest you to see links and attached files on topic.

Best regards

This depends on the aim of the pre-processing, you may need to illustrate your question.

I strongly recommend Matlab

（1）实在没有办法，可以用“枚举法”结合领域知识“试探”。

尽管“枚举法”很慢，但结果很可靠。例如，“枚举法”可以发现某些迭代法不收敛情况下的解，例如不稳定平衡点。

If really no way, please use "enumeration" combined with domain knowledge to "try".

Although the "enumeration" is inconceivable slow, the results are reliable. For example, the "enumeration method" can find solutions when some iterative methods do not converge, such as "unstable equilibrium points".

（2）求局部解，也可以使用“回归”。

To find some local solutions, may try to "regression".

Dear Wellington Pinheiro Dos Santos, I am now dealing with a book chapter dealing polymeric nanoparticles (PN) for solid cancer drug delivery. During my bibliographic processing I found appreciable bibliographic material on the same PN used in cancer diagnosis for image contrast mainly. However, in the recommended parts of your Book no one is concerned by this aspect. So, what do you think, if you think this is feasable I Can be in charge of that. Kindest Regards

Thank you so much for providing me with useful information.

Hi Daniel,

Yes, you can extract the same features from EMG/EEG signals.

On the other hand,

The first difference/divergence point is related to how you prepare your EMG/EEG signals for feature extraction. As you know, EEG is decomposed into several bands, these are: delta (<=4Hz), theta (4-12Hz), alpha (7-13Hz or 8-14Hz depending on source of publication), beta (14-30Hz), and gamma (>30Hz) (each with a certain frequency range that varies between publications). You can extract any time-domain, frequency-domain, or time-frequency domains features either from the original EEG signal (with full spectrum, say 0--58Hz or whatever range), or from the individual spectral bands (after filtering to get that band of interest). In comparison to EMG, we usually extract the features from one band (10Hz to nearly 500Hz). Or, you can also partition the EMG spectrum into ranges and extract features from each range just like EEG (you will need to determine the EMG ranges for your task accordingly).

The other divergence parameters could be related to how you further clean the signals, with EEG you need to clean eye-blinks and/or jaw movements away and with EMG you need to clean ECG beats (depending on where you collected the signals from). There could be other factors to worry about (depending on your application).

Other divergence points between the two signals' feature extraction process could also be related to the windows size/other parameters. For example, with EMG we usually use 100 msec to 150 msec windows when working on prosthesis control and may be larger values for different applications (usually small values depending on the application). On the EEG side, this will vary significantly depending on the application, for example if you are classifying sleep stages then a window size of 1 sec to 10 sec is fine, as you don't expect someone to go from fully awake to sleepy in less than 1 sec. Similarly for drowsiness detection (transition from awake to drowsy is gradual), an EEG window size of a few seconds has also been utilized. This is an application driven parameter.

I would suggest you start with simple time-domain features first, especially something like "Hjorth parameters". Hjorth parameters extract some frequency-domain characteristics directly from the time-domain without having to use any frequency analysis tools like FFT or wavelet etc.. I have used Hjorth on both EMG and EEG and got excellent results in many cases.

Hello, first of all, I would like to congratulate you on your new start, secondly, you should study survey and review papers on pattern recognition that will show you the existing methods and how the process should be also the same thing goes to the 3D reconstruction you should study image recomposition and stereoscopic images. Best of luck

I agree with Mark. I have also used Siamese. It works fine.

For more information:

With Matlab: https://www.mathworks.com/help/images/image-restoration-deblurring.html

For more information, we have:

https://www.hitechnectar.com/blogs/deep-learning-face-recognition/

I suggest you follow https://towardsdatascience.com/handling-overfitting-in-deep-learning-models-c760ee047c6e

https://elitedatascience.com/overfitting-in-machine-learning

https://www.kdnuggets.com/2019/12/5-techniques-prevent-overfitting-neural-networks.html

This link is important for your question:

Well, it is not that efficient solution but you can try to help the model by limiting the number of states, for instance, eliminiting use cases that you think are possible to take into consideration by the model but won't help you to reach your final objective. Or by limiting the action space, like the use of the grid world instead of continuous action space in case of navigation...

You will find in this link:

You could try with:

SAGA SIG :https://sourceforge.net/projects/saga-gis/

or Python:

ITKsnap, http://www.itksnap.org

3D Slicer, http://www.slicer.org/

ImageJ, http://imagej.nih.gov/ij/

Fiji (is ImageJ), http://fiji.sc/Fiji

TurtleSeg, http://www.turtleseg.org/

Dear Yosef Mousavi

Agree with many of the previous answers. It is based on the Fibronacci number or series. It is the infinite sequence of natural numbers:

0,1,1,2,3,5,8,13,21,34,55,89,144, ....

The golden ratio can be found in nature and is the shape of complex structures such as the petals of a flower (sunflower) or the leaves on a stem, or the branches of a tree, etc; They are organized in the smallest possible space while maintaining the most efficient architecture from a functional point of view.

regards

José Luis

Despite CNN and RNN, GCN is also helpful in handling many NLP applications for sequence learning incorporated with GANs structures. In my perspective, all deep learning-based networks either CNN, RNN, or GCN have an uniques impact on different applications in this scientific era. Yiwen Sun Reema Ahmad Muhammad Ali Adnan Riaz

The most important challenge is how to choose the suitable techniques for your problem

Yes, indeed. We need sequential models based on panel data and together with regression models we may need other statistical models based on computer vision related databases., before machine learning methods we need a lot of medical data as well numerical & visual data etc.

For this purpose, there is technical analysis, fundamental analysis, and scenario techno-fundamental automated system-cognitive analysis

There are actually quite a lot of nice application of machine learning techniques in the context of business process variant analysis, which is a fairly large subset of the process mining literature.

For example, Folino, Cuzzocrea et al. have done a series of studies on variant analysis (or deviance mining) using various machine learning methods, including ensemble learning and clustering:

We recently conducted a literature survey of methods in the field of variant analysis, many of them based on machine learning techniques:

Related to the above, there is work on bayesian networks for delay analysis (explanatory rather than predictive):

The above is related to variant analysis and performance mining. But there is also work on anomaly detection in event logs using bayesian networks:

And using deep learning architectures:

As well as using deep learning models to compute alignments in order to correct anomalies:

And a bit related to the above, there was quite a bit of research on using trace clustering in the context of automated process discovery (e.g. Jochen De Weerdt)

So we can say that process mining and machine learning go well together. One should not forget though that BPM and process mining are application-oriented disciplines - their objective is to design approaches to improve business processes. Whereas machine learning is a horizontal discipline, it seeks to develop methods that can be adapted to a broad range of problems/settings. Process mining has tapped a lot into machine learning, but sure it has a lot more to exploit from it.