Lamb Behaviors Analysis Using a Predictive CNN Model and a Single Camera

Round 1

Reviewer 1 Report

This manuscript proposed a CNN based model for predicting the behaviors of lamb. The idea is interesting and the result is basically satisfactory. However, some other problems in the manuscript are still concerned in the following:

1. Could the authors compare the proposed method with other state-of-the-art methods to validate the effectivity more extensively in the experiments?
2. I didn’t understand the sentence “In this study, he implements deep learning to detect the lamb’s postures (lying down, standing and eating) and basic computer vision techniques…”.
3. The language should be polished.
4. More technical details should be shown in Figure 1.

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 2 Report

This article provides a method to evaluate lamb behavior using a single RGB camera and a predictive model. First, a CNN-based lamb tracking architecture is conceived and implemented. Second, an animal behavior prediction model is proposed. The results show that the proposed methodology is viable and promising. YOLOV4 had a precision of 99.85 percent for detecting lamb activities, while the proposed prediction model had a precision of 83.52 percent for recognizing aberrant states. These findings show that the proposed methodology can be used in precision agriculture to detect diseases and problems.

In a satisfactory manner, the basic purpose of the research has been described, but with some crucial comments that should be taken into consideration.

1. In the instruction section, the related work needs some improvement as adding a table compare between the previous studies in terms of the year, Deep learning methods, the performance metrics, and its values.
2. In Section 3, Results, it is recommended to explain the results in terms of comparison with the previous studies that used deep learning methods by adding tables or suitable charts.
3. In the equations (12) and (13) check the symbols because in the first equation for Recall in the denominator the right is to write FN without (‘). The same for the precision equation; in the denominator, the right is to write FP without the upper dot.
4. There is only one reference from the year 2022 in the list of references. We recommend that the author include at least THREE current references that are from the year 2022 in order to improve the article.

   This article provides a method to evaluate lamb behavior using a single RGB camera and a predictive model. First, a CNN-based lamb tracking architecture is conceived and implemented. Second, an animal behavior prediction model is proposed. The results show that the proposed methodology is viable and promising. YOLOV4 had a precision of 99.85 percent for detecting lamb activities, while the proposed prediction model had a precision of 83.52 percent for recognizing aberrant states. These findings show that the proposed methodology can be used in precision agriculture to detect diseases and problems.

   In a satisfactory manner, the basic purpose of the research has been described, but with some crucial comments that should be taken into consideration.

   1. In the instruction section, the related work needs some improvement as adding a table compare between the previous studies in terms of the year, Deep learning methods, the performance metrics, and its values.
   2. In Section 3, Results, it is recommended to explain the results in terms of comparison with the previous studies that used deep learning methods by adding tables or suitable charts.
   3. In the equations (12) and (13) check the symbols because in the first equation for Recall in the denominator the right is to write FN without (‘). The same for the precision equation; in the denominator, the right is to write FP without the upper dot.
   4. Figure 5 is unclear; the author should replace it with a higher-quality image.
   5. There is only one reference from the year 2022 in the list of references. We recommend that the author include at least THREE current references that are from the year 2022 in order to improve the article.

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 3 Report

Authors implement deep learning to detect the lamb’s postures - lying down, standing and eating - and basic computer vision techniques to track objects correctly. They monitor the sheep, determine their activity, and identify their behaviour with video analysis. The method developed in this study proposes the automatic analysis of the feeding behaviour of fattening sheep and helps in precision farming in the application of preventive measures and diagnoses of possible ailments or problems.

Lack of the paper:

• the text has to be larger in fig. 2, 5, 8, 9 and tab. 1.
• why you don't use simply accuracy?
•  

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

I have no other concerns.

Reviewer 2 Report

All comments were taken into account. In its current state, the paper is sufficiently good as a stand-alone document to generate added value.