August 2021
·
49 Reads
·
12 Citations
What is this page?
This page lists works of an author who doesn't have a ResearchGate profile or hasn't added the works to their profile yet. It is automatically generated from public (personal) data to further our legitimate goal of comprehensive and accurate scientific recordkeeping. If you are this author and want this page removed, please let us know.
Publications (2)
![Fig 3. Architecture of Simple CNN This is a simple architecture which is provided for base comparison with other state-of-the-art models [3]. It uses the 2D Convolutional layers to capture the features of the images. Just the basic 3x3 filters can be used. Max pooling layers are important to reduce the dimensions of the input and the number of parameters to be learned. These Max pooling layers are added between the 2D convolutional layers. Then a layer is used to flatten the feature matrix to a column matrix. Then two fully connected layers are added. The cost function used in between the convolutional layers is RELU function to avoid the problem of vanishing gradients. The second type of cost function is used in the last fully connected layer which is soft max function which fits the cross-entropy loss function.](https://www.researchgate.net/publication/352380029/figure/fig1/AS:1034593918533639@1623677820755/Architecture-of-Simple-CNN-This-is-a-simple-architecture-which-is-provided-for-base_Q320.jpg)
February 2021
·
470 Reads
·
7 Citations
International Journal of Engineering Applied Sciences and Technology
As waste segregation becomes an important issue in our lives, with the use of technology like deep neural networks and computer vision, we can make the process efficient and robust by image segmentation and classification. These systems on the rise need accurate and efficient segmentation and recognition mechanisms and this demand coincides with the increase of computational capabilities of modern computer architectures and more effective algorithms for image recognition. This paper does a comparative analysis of various different approaches and methods like Simple CNN, ResNet50, VGG16, etc in brief. The comparative analysis and study explains the performance of every approach, this paper concludes that ResNet50 gives excellent performance. VGG16 network also provides good performance which meets the needs of daily use.
Citations (2)
... Augmentation techniques for addressing data scarcity are explored by 25 , while metadata-based approaches 26 and robotics-driven classification 27 contribute to the research landscape. Moreover, studies utilizing CNN models 28,29 underline the ongoing efforts to improve trash classification methodologies and their practical implications in waste management practices. Table 1 provides a comparison of some of the existing trash classification studies. ...
- Citing Conference Paper
August 2021
... Tiga teknik yang umum digunakan dalam domain ini adalah Histogram of Oriented Gradients (HOG), Local Binary Pattern (LBP), dan Zoning. HOG efektif dalam menangkap orientasi kontur serta informasi tepi dan gradien [16], [17], menjadikannya sangat cocok untuk mendeteksi arah objek [18]; LBP unggul dalam menggambarkan tekstur lokal melalui perbandingan intensitas piksel [19], [20], sehingga mampu mengatasi variasi pencahayaan dan transformasi [21], [22]; sementara Zoning membagi citra ke dalam grid untuk menghitung kepadatan piksel pada tiap zona [23], yang berkontribusi terhadap pemetaan spasial fitur. Kombinasi teknik-teknik ini dipercaya dapat memperkaya informasi fitur yang digunakan dalam pelatihan model klasifikasi [15], [22], [23] . ...
- Citing Article
- Full-text available
February 2021
International Journal of Engineering Applied Sciences and Technology