Camera Traps, Drones and Passive Acoustic Monitoring: Wildlife studies using recording devices to record still/video images and/or audio have the potential to capture human subjects. This has implications for privacy, and — in the event that illegal acts are captured — a duty on the part of the researcher to report them. There might also be issues of personal safety in the latter case. See https://doi.org/10.1002/2688-8319.12033 and https://doi.org/10.1111/csp2.374 for more information.
Ideally, this should have been considered by the authors and their ethics committee before the study was approved but we understand that this is not currently a widespread practice. While the field is reaching a consensus on this matter, we ask authors who are reporting data from such studies to explain how the ethical approval/field permit they obtained relate to these issues, or explain why this was not considered necessary.
Where the study uses mobile platforms such as uncrewed aerial vehicles (UAVs/drones) these must have been operated in accordance with all applicable regulations.
This policy is likely to adapt as the field develops.
PeerJ Computer Science is committed to improving scholarly communications and as part of this commitment, authors must make materials, code, data, and associated protocols available at the time of submission for peer review and publication. The preferred way to meet this requirement is to publicly deposit as noted below. Cases of non-compliance will be investigated by PeerJ Computer Science which reserves the right to act on the results of the investigation.
We strongly recommend (and in some cases require) that authors adhere to the reporting standards which have been adopted by their field (or which apply to their study design).
All statistical results should be reported in full, including the test that was performed, the reason for choosing that test, the corresponding test statistic, sample size, degrees of freedom, the exact p-value expressed up to 2 decimal spaces unless 'p<0.001' or confidence interval, and effect sizes. Where multiple testing is performed, suitable corrections must be made.
Do not report inferential statistics such as p values or confidence intervals for known quantities such as baseline measurements. The spread of the data can be indicated by descriptive statistics such as standard deviation, or quantiles and ranges.
Where appropriate, we recommend that you overlay bar graphs with scatter plots showing individual data points, or use another method to show the distribution of the data, such as boxplots, violin plots, etc.
PeerJ journals consider timely and well-balanced literature reviews of fields with broad cross-disciplinary interest within the journal's scope. While we do not impose a hard limit, we recommend a maximum of 8,000-12,000 words in order to keep the review focused. The review should include a rationale for why it is needed, describe who it is intended for, and include a description of the procedures used to ensure that it is comprehensive and unbiased (for example, the search strategies that were employed). Gaps in the literature, future avenues of research and opportunities for cross-disciplinary collaborations should be clearly identified. Unbalanced reviews that are performed with the intention of supporting a particular interpretation or point of view will not be considered.
Since, by their nature, literature reviews rely heavily on the published work of others, it is especially important to avoid inadvertent plagiarism by copying and pasting sections of text from the original source. In addition, it is very important, when quoting or paraphrasing, to correctly acknowledge your sources.
We recommend that your review is structured following the guidelines for Literature Review Articles in standard sections.
AI Application articles present research relating to the study of artificial intelligence techniques (for instance, machine learning, deep learning, algorithms, computer vision, natural language processing, and intelligent systems) and the application of these techniques to areas that need not necessarily be related to computer science (e.g. medicine, life sciences, social sciences, geographic sciences, chemistry, education, business, etc.). Simple application (or cross-evaluation) of existing tools to datasets without a clear articulation of the need for, and limitations of, the method(s) will not be considered.
AI Application articles must describe the need for the technique, and demonstrate how the machine learning techniques discussed can be used to solve practical problems. The article must include a research methodology that demonstrates improvement to current practices in the relevant field. Where possible, the technique should be compared to existing methods, to demonstrate the improvement.
Submitted articles must be technically robust and clearly presented; code and data (original or third party) must also be freely available to ensure sound science and reproducibility.
In the interests of inclusivity, PeerJ does not condone the use of images such as Lena/Lenna and Tiffany without a strong scientific justification. Suitable substitutes are available. If you must use the image to compare the performance of your algorithm to a published paper where the algorithm is not available, please avoid reproducing the image in your figures and only report the numerical results (e.g. histograms, etc.).
In general, it is expected that work involving human subjects will be submitted to PeerJ - the Journal of Life and Environmental Science, but if the intended audience is the computer science community and the human experiments are a minor part of the overall work then it will be considered for publication in PeerJ Computer Science, so long as it complies with the usual ethical requirements.
These policies are made available under the Creative Commons CC BY 4.0 license and can be copied for reuse with attribution.