Skip to content

Latest commit

 

History

History
28 lines (23 loc) · 6.3 KB

README.md

File metadata and controls

28 lines (23 loc) · 6.3 KB
Raw

cakalang

Abstrak

Upaya menjaga kelestarian usaha perikanan tangkap terutama yang menghasilkan devisa yaitu ikan Tuna, Tongkol dan Cakalang perlu dilakukan untuk menghindari kondisi overfishing yang akan berakibat pada kepunahan. Kontribusi penelitian dan novelti yang dicapai adalah percepatan pada proses yang terdapat pada algoritma Maxpool. Pembuatan sistem yang dapat mengenali Katsuwonus Pelamis (Ikan Cakalang), Euthynnus Affinis (Ikan Tongkol), Coryphaena Hippurus (Ikan Lemadang) dan Loligo Chinensis (Cumi-cumi), mendeteksi panjang ikan dan bobotnya dengan menggunakan kamera, yang sebelumnya dilakukan pengukuran dan penimbangan ikan secara manual dengan latar belakang yang berbeda-beda dan jarak yang statis. Penelitian juga melakukan modifikasi maxpool layer sehingga dapat meningkatkan kinerjanya. Model yang dilatih dapat melakukan pendeteksian obyek dengan tingkat akurasi 89.55%. Penelitian menghasilkan error estimasi panjang sebesar 2.59%, error estimasi berat sebesar 15.67%, dan percepatan maxpool 221.01 ms (27.99%).

Efforts to maintain the sustainability of capture fisheries, especially those that generate foreign exchange, namely Tuna, Tongkol and Skipjack tuna need to be carried out to avoid overfishing conditions that will result in extinction. The research contribution and the novelty achieved is the acceleration of the process contained in the Maxpool algorithm. Making a system that can recognize Katsuwonus pelamis (Squid Skipjack), Euthynnus Affinis (Tub Fish), Coryphaena Hippurus (Lemadang Fish) and Loligo Chinensis (Squid), detect fish length and weight using a camera, previously measured and weighed fish manually with different backgrounds and static spacing. The research also modifies the maxpool layer so that it can improve its performance. The trained model can detect objects with an accuracy rate of 89.55%. The study resulted in a length estimation error of 2.59%, a weight estimation error of 15.67%, and a maxpool acceleration of 221.01 ms (27.99%).

Papers

  1. Liawatimena, S., Heryadi, Y., Trisetyarso, A., Wibowo, A., Abbas, B. S., & Barlian, E. (2018, September). A fish classification on images using transfer learning and matlab. In 2018 Indonesian association for pattern recognition international conference (INAPR) (pp. 108-112). IEEE.
  2. Liawatimena, S., Atmadja, W., Abbas, B. S., Trisetyarso, A., Wibowo, A., Barlian, E., ... & Zulardi, I. (2020, February). Computer Vision and Fuzzy Logic for Sustainable Indonesian Fisheries. In IOP Conference Series: Earth and Environmental Science (Vol. 426, No. 1, p. 012154). IOP Publishing.
  3. Liawatimena, S., Atmadja, W., Abbas, B. S., Trisetyarso, A., Wibowo, A., Barlian, E., ... & Yojana, A. C. (2020, February). Drones Computer Vision using Deep Learning to Support Fishing Management in Indonesia. In IOP Conference Series: Earth and Environmental Science (Vol. 426, No. 1, p. 012155). IOP Publishing.
  4. Liawatimena, S., Abdurahman, E., Trisetyarso, A., Wibowo, A., Atmadja, W., Effendi, F., & Edbert, I. S. (2021, July). Convolve4D: A Novelty Approach to Improve Convolutional Process. In IOP Conference Series: Earth and Environmental Science (Vol. 794, No. 1, p. 012107). IOP Publishing.
  5. Liawatimena, S., Abdurahman, E., Trisetyarso, A., Wibowo, A., Edbert, I.S., Ario, M.K., & Effendi, F. (July 2022). Performance Optimization Of Maxpool Calculation Using 4D Rank Tensor. In ICIC Express Letters, Part B: Applications (Vol. 13, No. 7, pp767-776). Tokai University, Japan
  6. Liawatimena, S., Abdurachman, E., Trisetyarso, A., Wibowo, A, Ario, M.A., Edbert, I.S. (March 2023). Fish Classification System Using YOLOv3-ResNet18 Model for Mobile Phones. In CommIT (Communication and Information Technology) Journal (Vol. 17, No. 1)
  7. Liawatimena, S. (2022). Akselerasi Maxpool Pada Pengenalan Obyek dan Pengukuran Panjang Katsuwonus Pelamis, Euthynnus Affinis, Coryphaena Hippurus dan Loligo Chinensis Menggunakan Convolutional Neural Network. Disertasi Doktoral, Universitas Bina Nusantara.
  8. Presentasi
  9. Poster
    Poster
  10. Video Sambutan oleh Dr. Lukas (Ketua Indonesia Artificial Intelligence Society)
    Video Sambutan oleh Dr. Lukas (Ketua Indonesia Artificial Intelligence Society)
  11. Video A Fish Detection Using ExeML ModelArts
    Watch the video

Code:

  1. Convolve4D (Python), related with paper #4
  2. Maxpool (Python), related with paper #5

Dataset:

  1. 4000 images + annotations

Bila Anda menggunakan dataset atau kode kami, mohon paper-paper kami bisa dicantumkan dalam referensi penelitian Anda. Terima kasih banyak sebelumnya. If you use our dataset or code, please include our papers in your research reference. Thank you in advance.