ACOGARE: Acoustic-Based Litter Garbage Recognition Utilizing Smartwatch
Round 1
Reviewer 1 Report
An interesting idea proposed by the paper, but in its current state, it must be worked on.
i) Formalize the detection of the type of Garbage. (note: not to be confused with some kind of non-junk and/or non-game object)
ii) How would the modulation and demodulation model be in the detection of littering by a person?
iii) Test scenarios should be presented and compared to existing systems, such as littering and no littering.
iv) Justify why the three ML methods proposed for the study were chosen.
v) The evidence is poor to support the proposed conclusions.
Best Regards
Moderate editing of English language
Author Response
Please see the attachment
Author Response File: 
Author Response.docx
Reviewer 2 Report
Figure 8,9 and 10 should be moved as they are placed in the references section making them hard to find. The results of the study would be even better it there would have been more data for the ML models.
The completed work is comprehensive and interesting for readers, especially for those who are new to this field.
Although there is a lack of test data for the elements presented, the authors manage to reproduce the effects quite well and support the research. The work has potential for development and I believe that the problems addressed should be deepened.
Author Response
Please see the attachment
Author Response File: Author Response.docx
Reviewer 3 Report
In this work, the authors propose a conceptual design to estimate the types and locations of litter using only the sensor data from a smartwatch worn by the user to overcome the litter problem and plan for garbage bin placements. In addition, an acoustic recognition model was used to estimate the types of litters.
- The paper is well-structured and informative.
- The authors should consider discussing related deep learning models to tasks such as human activity recognition and signal processing (10.1016/j.future.2023.01.006 and 10.1016/j.measurement.2022.111445).
- The authors can improve their methodology and experiments by incorporating deep learning approaches.
- The conclusion looks consistent.
Author Response
Please see the attachment.
Author Response File: Author Response.docx
Reviewer 4 Report
The authors propose an acoustic based approach leveraging smartwatch for litter garbage recognition.
1. The authors propose a machine learning approach based on the 251 features. Does the deep learning approach work better? It would be interesting to investigate or compare with the deep learning approach.
2. I am still not clear about the motivation of litter garbage recognition base on smartwatch. Please elaborate on the motivation in the part of introduction.
3. Does the different activities or gender to pick up litter affect the recognition accuracy?
4. How about the computation time? Can the model be applied in the smartphone?
Author Response
Please see the attachment.
Author Response File: Author Response.docx
Round 2
Reviewer 1 Report
In this paper, they propose a conceptual design to estimate the types and locations of litter using only the sensor data from a smartwatch worn by a litter picker.
i) It must be improved. The description, specifications and design of the proposed model in the theoric/formal way. It must be clear and detailed in all its phases, from the pre-processing of the audio to the correct detection of the type of litter to be recognized.
ii) Related research should include more articles and related works.
iii) Organize in a better way sections 5 and 6.
best regards
What is the importance or explanation of using the word "litter" instead of: garbage or trash?
Author Response
Please see the attachment.
Author Response File: 
Author Response.pdf
Reviewer 3 Report
The authors addressed all my concerns.
Author Response
Thank you for your comments. I would like to continue working on my research.
Round 3
Reviewer 1 Report
Accept in present form
Best Regards
Minor editing of English language required
Best Regards
