search

ZhuoYulang / IF-MMIN

MIT License
28 stars 3 forks source link

May I ask if there have been any studies or research conducted on the IEMOCAP dataset involving a 6-class classification task? #1

Open mirrorballty opened 11 months ago

mirrorballty commented 11 months ago

May I ask if there have been any studies or research conducted on the IEMOCAP dataset involving a 6-class classification task?

ZhuoYulang commented 9 months ago

I apologize for not getting back to you in time. Currently we do not have experiments on the IEMOCAP involving a 6-class. Best Wishes---- Replied Message @.>Date11/27/2023 @.>@.>Subject[ZhuoYulang/IF-MMIN] May I ask if there have been any studies or research conducted on the IEMOCAP dataset involving a 6-class classification task? (Issue #1)May I ask if there have been any studies or research conducted on the IEMOCAP dataset involving a 6-class classification task?—Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you are subscribed to this thread.Message ID: @.>

mirrorballty commented 6 months ago

Hello, I recently downloaded the feature sets for the two datasets provided through the Baidu Cloud link. I'm particularly interested in understanding the nature of the feature extraction process for the three modalities. Could you kindly confirm if these features were obtained through a Word-aligned setting or an Unaligned setting? Clarity on this aspect will significantly aid in my analysis. Thank you for your time and assistance.

ZhuoYulang commented 6 months ago

We used the features and framework provided by MMIN. And we did not additionally consider the word-aligned setting during our experiments. ---- Replied Message ---- @.>Date5/11/2024 @.>@.>, @.>SubjectRe: [ZhuoYulang/IF-MMIN] May I ask if there have been any studies or research conducted on the IEMOCAP dataset involving a 6-class classification task? (Issue #1) I was wondering if you conducted your experiments in a Word-aligned setting or an Unaligned setting. — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.Message ID: @.***>

ZhuoYulang commented 6 months ago

Hello, This is the paper link of MMIN --- Missing Modality Imagination Network for Emotion Recognition with Uncertain Missing Modalities: https://aclanthology.org/2021.acl-long.203/ And our work is based on MMIN, including dataset. From paper MMIN, we don't find the word-aligned process, but we noted that the feature extracted processes of three modalities are individual. And there is no related module to align modality features in MMIN. So we think the features of these modalities are unaligned. So, in our model, we similarly did not consider modality alignment. And the features that from the MMIN also are unaligned. You can search more details in the MMIN paper. I hope these can help you. ---- Replied Message ---- @.>Date5/11/2024 @.>@.>, @.>SubjectRe: [ZhuoYulang/IF-MMIN] May I ask if there have been any studies or research conducted on the IEMOCAP dataset involving a 6-class classification task? (Issue #1) Hello, I recently downloaded the feature sets for the two datasets provided through the Baidu Cloud link. I'm particularly interested in understanding the nature of the feature extraction process for the three modalities. Could you kindly confirm if these features were obtained through a Word-aligned setting or an Unaligned setting? Clarity on this aspect will significantly aid in my analysis. Thank you for your time and assistance. — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.Message ID: @.***>

mirrorballty commented 6 months ago

Thank you for your response. My research focus is on multimodal emotion recognition involving speech, text, and video. However, I am not familiar with feature extraction from these modalities. Therefore, I downloaded the feature sets of these two datasets from the Baidu Cloud link provided in the MMIN paper. Based on these features, I conducted my research, which is similar to what you refer to as full multimodal emotion recognition in your work. In addition, while going through other papers, I came across two scenarios: Word-aligned setting and Unaligned setting. Moreover, I noticed the presence of the "make_aligned.py" file in the IF-MMIN/preprocess/IEMOCAP/ directory of the provided code repository. This led me to speculate if it is somehow related to the Word-aligned setting, hence my previous question.

ZhuoYulang commented 6 months ago

Yeah, maybe you are right. I skip the feature extracted process so I ignore the files in IF-MMIN/preprocess/IEMOCAP/. I'm sorry I for almost misleading you. You can connect with the author of MMIN and find more details. Good luck with your research! ---- Replied Message ---- @.>Date5/11/2024 @.>@.>, @.>SubjectRe: [ZhuoYulang/IF-MMIN] May I ask if there have been any studies or research conducted on the IEMOCAP dataset involving a 6-class classification task? (Issue #1) Thank you for your response. My research focus is on multimodal emotion recognition involving speech, text, and video. However, I am not familiar with feature extraction from these modalities. Therefore, I downloaded the feature sets of these two datasets from the Baidu Cloud link provided in the MMIN paper. Based on these features, I conducted my research, which is similar to what you refer to as full multimodal emotion recognition in your work. In addition, while going through other papers, I came across two scenarios: Word-aligned setting and Unaligned setting. Moreover, I noticed the presence of the "make_aligned.py" file in the IF-MMIN/preprocess/IEMOCAP/ directory of the provided code repository. This led me to speculate if it is somehow related to the Word-aligned setting, hence my previous question. — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.Message ID: @.***>

mirrorballty commented 6 months ago

I apologize for the inconvenience caused and I'm sorry to bother you. Could I kindly request to exchange contact information for further communication? My email address is 2387380534@qq.com.

ZhuoYulang commented 6 months ago

Of course. My email is @.  ---- Replied Message ---- @.>Date5/11/2024 @.>@.>, @.>SubjectRe: [ZhuoYulang/IF-MMIN] May I ask if there have been any studies or research conducted on the IEMOCAP dataset involving a 6-class classification task? (Issue #1) I apologize for the inconvenience caused and I'm sorry to bother you. Could I kindly request to exchange contact information for further communication? My email address is @. — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.Message ID: @.***>

mirrorballty commented 6 months ago

I cannot see your email information. You can directly send an email to 2387380534@qq.com.