The Indian Deepfaker, has recently collaborated with Twilight Entertainment and Wakefit, a renowned mattress brand, to create a groundbreaking advertisement featuring the talented Bollywood actor Ayushman Khurrana. 🎥💫
This is the first Deepfake + deaging ad created by an Indian company.
In the fast-paced world of advertising, capturing audience attention and standing out from the competition is a constant challenge.
At The Indian Deepfaker, we excel at pushing boundaries and utilizing cutting-edge technology to create captivating experiences.
Our latest triumph? A deepfake ad crafted in collaboration with Twilight Entertainment and Wakefit, featuring the talented Bollywood actor Ayushmann Khurrana.
In this blog post, we invite you on a fascinating journey as we decode the magic behind the creation of this revolutionary ad, showcasing the immense power of artificial intelligence (AI).
Before we dive into the details, we invite you to watch the captivating video that takes you behind the scenes of our deepfake ad creation.
Get ready to witness the fusion of technology, creativity, and talent, as Ayushmann Khurrana’s transformation unfolds before your eyes.
Harnessing the Power of AI:
Creating a deepfake ad is no small feat. It requires a deep understanding of AI techniques and an eye for detail. Our journey began with training a sophisticated AI model using Generative Adversarial Networks (GANs). This process allowed us to capture Ayushmann Khurrana’s facial features and seamlessly superimpose them onto the body of a child actor, creating a jaw-dropping transformation that defies reality.
Decoding the AD:
Step 1: Data Collection
Our project involved swapping the face of actor Ayushman Khuranna with a child in the Wakefit ad. In the ad, Ayushman transforms into a child version of himself after waking up. To achieve this, we needed to address two challenges.
First, we used deepfake technology to swap Ayushman Khuranna’s face with the child in the ad. Deepfake technology allows us to manipulate and replace faces in videos. To gather the necessary data, we collected footage of Ayushman from one of his movies, “Bala,” where he played a bald character. In most of Ayushman Khuranna’s interviews and movies, he either wears goggles or has a beard. However, for high-quality results, we needed clean-shaven footage without goggles. We extracted video segments from “Bala” where Ayushman’s face was clearly visible in different lighting conditions and environments. We also obtained the destination video where the face swap was required.
Step 2: Video to Frames Conversion
Once we had the video data, we ran an algorithm to convert the video into individual frames. This allowed us to process the frames individually and manipulate them as needed.
Step 3: Face Extraction
We then used another algorithm to extract the faces from the frames. By isolating the faces, we could focus on the specific area that required modifications.
Step 4: Data Cleaning
To ensure high-quality results, we cleaned the data by removing frames where no face was detected or the face was blurred. The quality of the training data directly affects the final outcome, so we only kept frames with clear and identifiable faces.
Step 5: Facial LandMarks Detection
Next, we utilized an algorithm to detect facial features and landmarks within the frames. This step helped in aligning the faces properly and maintaining consistency across the frames.
Step 6: Training the Model
We trained the model in 3 different part with different parameter settings using GAN’S framework.
We proceeded by employing a GAN (Generative Adversarial Network) framework, which utilizes artificial neural networks. The GAN framework consists of two AI models: a Generator and a Discriminator. The generator generates images from noise, while the discriminator evaluates whether the generated image resembles the face that needs to be swapped.
The models undergo training iterations until the discriminator can no longer distinguish between real and fake images. We trained the model for 1 million iterations to achieve the desired results. Once trained, we used another algorithm to merge the generated faces.
Step: 7 Editing and Fine Tuning
After merging the trained faces, we performed further editing and fine-tuning using After Effects, a software for video post-production. This allowed us to refine the deepfake video and ensure its seamless integration.
Step:8 The Art of Deaging Face
Apart from swapping faces, our second challenge was to deage Ayushman Khurana’s face to resemble his childhood version. Deaging involves making someone appear younger using AI techniques. However, we faced a scarcity of Ayushman’s childhood photos, with only two images available online.
Nevertheless, our team worked diligently to find a solution. We utilized machine learning algorithms to deage each frame of Ayushman’s source data and retrained the model using the same destination video. After final fine-tuning, we successfully achieved the desired deaging effect.
Unlike a foreign company’s similar ad featuring a deaged Sachin Tendulkar, who had plenty of childhood data due to his cricket career, our limited resources presented a challenge.
By following these steps and overcoming the challenges, we were able to create the Wakefit ad deepfake video starring Ayushman Khurrana, where his face is swapped with a child actor and deaged to resemble a younger version of himself.
The result? A child version of Ayushmann that bears an uncanny resemblance to the actor, adding an extra layer of nostalgia and charm to the ad.
Collaboration and Creative Vision:
The success of this groundbreaking project would not have been possible without the collaboration and creative vision @twilight entertainment, @sushant baliga ,Prateek Malpani , Wakefit and Ayushman Khurrana. Their enthusiasm and openness to innovative ideas allowed us to embark on this uncharted territory in the advertising landscape.
Together, we shattered conventions and crafted an ad that leaves a lasting impression, capturing the essence of Wakefit’s brand and the transformative power of a good mattress.
Setting New Standards:
As the first Indian company to accomplish such a feat, we take immense pride in our groundbreaking achievement.
This deepfake ad not only breaks through the clutter but also sets new standards for creative storytelling and brand communication.
It showcases the transformative potential of AI and deepfake technology, redefining how brands can captivate and engage audiences in an increasingly competitive advertising landscape.
Conclusion:
The deepfake ad created by The Indian Deepfaker for Wakefit stands as a testament to our commitment to excellence, innovation, and the power of AI.
Through this project, we have unveiled the artistry and magic of deepfakes, seamlessly merging talent and technology to deliver a truly captivating advertisement.
Join us in celebrating this remarkable achievement and let’s continue pushing the boundaries of what is possible in the world of advertising.
Discover more about our deepfake and deaging services and explore the limitless possibilities for your brand.
